---
title: "GPT (OpenAI)"
source: "https://systems-analysis.info/eng/GPT_(OpenAI)"
wiki: "systems-analysis.info/eng"
article: "GPT_(OpenAI)"
language: "en"
categories:
  - "Category:English"
  - "Category:Large language models"
  - "Category:LLM families"
  - "Category:Machine learning"
  - "Category:OpenAI"
  - "Category:Technology"
revision_id: 160
wiki_created_at: 2026-09-06T22:18:13Z
wiki_modified_at: 2026-09-06T22:18:13Z
downloaded_at: 2026-09-07T22:21:27Z
---

# GPT (OpenAI)

**GPT** (*Generative [Pre-trained](https://systems-analysis.info/eng/Pre-training_of_large_language_models "Pre-training of large language models") Transformer*) is a family of [large language models](https://systems-analysis.info/eng/Large_language_model "Large language model") (LLMs) developed by OpenAI. GPT models are built on the [transformer architecture](https://systems-analysis.info/eng/Transformer_architecture "Transformer architecture") and implement the generative pre-training paradigm: in the first stage, the model is trained on extensive text corpora without explicit labeling, and can then be [fine-tuned](https://systems-analysis.info/eng/Fine-tuning_(deep_learning) "Fine-tuning (deep learning)") for specific tasks. For later generations (starting with GPT‑5), OpenAI also uses the term **unified system**, as the product combines a fast response mode, a deep reasoning mode, and a router<sup>[\[1\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt5-announce-1)</sup>.

## Name

The abbreviation **GPT** stands for **Generative Pre-trained Transformer**.

- **Generative**: indicates that the model is capable of creating (generating) new content, such as text.
- **Pre-trained**: **indicates that the model undergoes an extensive initial training stage on a large dataset (e.g., texts from the internet). After pre-training, the model can often be additionally "fine-tuned" for more specific tasks.**
- **Transformer**: **refers to a specific neural network architecture that is a key innovation underlying GPT and many other modern AI models.**

The main characteristic of GPT is that training occurs in an autoregressive manner — the model predicts the next [token](https://systems-analysis.info/eng/Token_(LLM) "Token (LLM)") based on the preceding context. That is, the model is trained to maximize the probability of the next token given a sequence of previous tokens. During training, the prediction error for the next element is minimized, which enables the generation of texts with high <a href="https://en.wikipedia.org/wiki/Coherence_(linguistics)" class="external text" rel="nofollow">coherence</a> and consistency.

## Text generation process in GPT

The GPT model generates text sequentially, token by token, according to the following iterative scheme:

- Receives an initial text sequence (prompt, seed text) as input.
- Computes a probability distribution over all tokens in the vocabulary for the next text element.
- Selects the next token:
  - either by the highest probability (greedy selection),
  - or by stochastic sampling (*sampling*),
  - or using special filtering strategies (*[top-k](https://systems-analysis.info/eng/Top-k_sampling "Top-k sampling")*, *[top-p](https://systems-analysis.info/eng/Top-p_sampling "Top-p sampling")*).
- Appends the selected token to the current sequence.
- The updated sequence is again fed into the model for predicting the next token.

## Transformer architecture: text processing

The data processing inside the ***transformer*** for predicting the next token involves several main stages:

- **[Tokenization](https://systems-analysis.info/eng/Tokenization_(NLP) "Tokenization (NLP)") (Tokenization).** The input text is split into tokens — small text units that can be words, subwords, or punctuation marks. In GPT-3, for example, the vocabulary includes approximately 50,257 tokens.
- **Token [embeddings](https://systems-analysis.info/eng/Embedding_(NLP) "Embedding (NLP)") (Embeddings).** Each token is converted into a fixed-length vector using an embedding matrix (W_E). The vectors encode token meanings: semantically similar tokens are located close together in the high-dimensional space. In GPT-3, the embedding dimensionality is 12,288.
- **Processing in transformer layers.**
  - **Attention blocks:** Each token interacts with other tokens in the sequence. The attention mechanism allows the model to account for context and correctly interpret word meanings.
  - **Feed-forward layers:** After attention, each token is processed individually through a two-layer neural network with nonlinear activation.
- **Reverse transformation and Softmax.** After all layers, the processed vector is transformed back into token space using a matrix (W_U), which is often a transposed version of W_E. The resulting logits vector is normalized using the Softmax function to obtain a probability distribution over all tokens.
- **Next token selection (Sampling).** The next token is selected based on the probability distribution. The [temperature](https://systems-analysis.info/eng/Temperature_(LLM) "Temperature (LLM)") parameter controls the randomness of selection: at temperature 0, the most probable token is selected; at higher temperatures, the probability of selecting less likely options increases, which promotes greater text diversity.

## GPT models

- **GPT-1 (2018):** the first model in the family; a 12-layer [decoder-only transformer](https://systems-analysis.info/eng/Decoder-only_models_(architecture) "Decoder-only models (architecture)"); two-stage training (pre-training + fine-tuning on NLP tasks).
- **GPT-2 (2019):** 1.5 billion parameters; trained on the WebText corpus; the first model capable of generating long coherent texts; improved zero-shot generation quality. Announced on February 14, 2019; the full version (1.5B) was released on November 5, 2019 due to safety concerns.
- **GPT-3 (2020):** 175 billion parameters; large-scale training on a combination of Common Crawl, Books, and Wikipedia; strong development of few-shot and zero-shot capabilities.
- **GPT-3.5 (2022):** an intermediate version between GPT-3 and GPT-4; improved instruction following through [Reinforcement Learning from Human Feedback](https://systems-analysis.info/eng/Reinforcement_learning_from_human_feedback_(RLHF) "Reinforcement learning from human feedback (RLHF)") (RLHF) in the text-davinci-003 and gpt-3.5-turbo versions; [context window](https://systems-analysis.info/eng/Context_window "Context window") up to 4,096 tokens in early versions and up to 16,385 tokens in later ones (gpt-3.5-turbo-16k and updated gpt-3.5-turbo).
- **GPT-4 (2023):** a [multimodal](https://systems-analysis.info/eng/Multimodal_large_language_models "Multimodal large language models") model with text and image input (image support was deployed later, after the text-only launch); context window of 8,192 tokens in the base version and 32,768 tokens in the GPT-4-32k variant; significant improvements in accuracy, robustness, and reasoning.
- **GPT-4 Turbo (2023):** an optimized version of GPT-4; increased context window up to 128,000 tokens; lower latency and cost.
- **GPT-4o (2024):** a next-generation multimodal model (text, image, audio) with a unified neural network architecture; very high response speed and accuracy; context window of 128,000 tokens.
- **GPT-4.5 (2025):** a research preview; the OpenAI system card states that the model "builds on GPT-4o"<sup>[\[2\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-2)[\[3\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt45-systemcard-3)</sup>; improved understanding of user queries, reduced error rate; context window of 128,000 tokens. The API model `gpt-4.5-preview` was declared deprecated on April 14, 2025 and shut down on July 14, 2025<sup>[\[4\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-deprecations-4)</sup>.
- **GPT-4.1 (2025):** an improved version of the GPT-4 family with a context window of up to 1 million tokens; accepts text and images as input, outputs text<sup>[\[5\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-5)</sup>. Released simultaneously in three variants: GPT-4.1, GPT-4.1 mini, GPT-4.1 nano.
- **GPT-5 (2025):** a unified system with fast response and deep reasoning modes; context window of approximately 400,000 tokens; notable reduction in [hallucinations](https://systems-analysis.info/eng/LLM_hallucinations "LLM hallucinations") on factual tasks.
- **GPT-5.1 (2025):** adaptive reasoning, improvements in coding and long-context retention.
- **GPT-5.2 (2025):** focus on professional work; Pro mode for frontier tasks; the agentic GPT-5.2-Codex was released based on GPT-5.2.
- **GPT-5.3-Codex (2026):** an agentic coding model combining coding capabilities and reasoning; 25% faster than predecessors.
- **GPT-5.3 Instant (2026):** an update to the most widely used conversational model in ChatGPT; released on March 3, 2026. Improved factual accuracy, web search quality, conversational flow, and reduced excessive refusals and unnecessary caveats. Available in the API as `gpt-5.3-chat-latest`<sup>[\[6\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt53-instant-6)</sup>.
- **GPT-5.4 (2026):** OpenAI's frontier model for professional work, introduced on March 5, 2026; the first general-purpose OpenAI model with native computer-use capabilities. In the API, `gpt-5.4` is recommended as the default model for a wide range of general-purpose and coding tasks<sup>[\[7\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt54-announce-7)[\[8\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt54-api-8)</sup>.

## GPT-1

The first model, **GPT-1**, was introduced by OpenAI in 2018 in the paper "*Improving Language Understanding by Generative Pre-Training*". The model was a 12-layer decoder-only transformer<sup>[\[9\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-radford2018-9)</sup> built on the transformer architecture. GPT-1 training proceeded in two stages: an unsupervised generative pre-training stage (*pre-training*), followed by a supervised fine-tuning stage (*fine-tuning*).

During the pre-training stage, the model was trained on the BookCorpus, comprising over 7,000 unpublished books of various genres. A distinctive feature of this corpus was the presence of long continuous text passages, which was critically important for developing the model's ability to process complex and long-range textual dependencies.

During the fine-tuning stage, the model was adapted to solve specialized natural language processing tasks, including:

- Question Answering (QA) — generating a correct answer based on a given textual context;
- Natural Language Inference (NLI) — determining the logical relationship between two texts: entailment, contradiction, or neutrality;
- Semantic Textual Similarity — measuring the degree of semantic closeness between two text sequences.

Thanks to this approach, GPT-1 demonstrated significant superiority over previous models on a number of standard benchmarks for text comprehension tasks.

The development of GPT-1 demonstrated several key achievements and discoveries in natural language processing (NLP):

- **Effectiveness of generative pre-training.** It was empirically confirmed that pre-training on large corpora of unlabeled text enables the model to acquire universal language representations suitable for subsequent application in various tasks without requiring fundamental architectural changes.
- **Versatility of the transformer architecture.** The use of a multi-layer decoder transformer enabled the model to successfully process long-range dependencies in text, which had previously been difficult for models based on recurrent neural networks.
- **Reduced dependence on labeled data.** The work confirmed that large-scale pre-training on unlabeled data can significantly reduce the amount of labeled data needed to achieve high quality on target tasks.
- **Foundation for further development.** The results of GPT-1 laid the conceptual and technical groundwork for subsequent versions of the GPT family (GPT-2, GPT-3, and beyond).

## GPT-2

The GPT-2 model was announced by OpenAI on February 14, 2019. It significantly surpassed its predecessor in size: the full version of the model contained approximately 1.5 billion parameters. For safety reasons, OpenAI initially released only smaller variants of the model; the full version (1.5B parameters) was released on November 5, 2019. Unlike GPT-1, which was trained on the BookCorpus (~5 GB), GPT-2 was trained on a specially compiled WebText corpus of approximately 40 GB, comprising textual data from high-quality internet sources. The increase in both model size and training data volume enabled GPT-2 to significantly improve text generation quality: it demonstrated the ability to create substantive articles, stories, and even coherent passages of fiction.

GPT-2 employed an autoregressive decoder-only transformer architecture similar to GPT-1, without significant changes. The model consisted of 48 self-attention layers, had a hidden state size of 1,600, and included approximately 1.5 billion parameters. The number of attention heads was 25 (maintaining a head size of 64, inherited from GPT-1: 1,600 ÷ 64 = 25). Training was performed on the next-token prediction task based on the preceding context using masked attention.

One of the main distinctions of GPT-2 was that the model was the first to demonstrate high effectiveness in **zero-shot learning** — the ability to solve new tasks without undergoing explicit fine-tuning on examples for those tasks. The model was trained on a large corpus of general texts and did not undergo specialized training on task-specific data. Evaluation was conducted in a **zero-shot** regime, in which the model performed tasks solely based on knowledge acquired during pre-training. On a number of language modeling tasks, GPT-2 achieved quality comparable to or exceeding the results of models specifically trained on specialized datasets (e.g., Wikipedia, news texts, books).

## GPT-3

The **GPT-3** model was introduced by OpenAI in June 2020 (the arXiv paper appeared on May 28, 2020; beta API access opened on June 11, 2020). It was the next step in the development of generative transformers after GPT-2 and was distinguished by scaling the architecture to 175 billion parameters, making it the largest language model at the time.

The architecture of GPT-3 remained fundamentally the same — a multi-layer autoregressive decoder-only transformer without radical changes. The main performance improvements were achieved through increasing the number of layers, the width of hidden layers, and the scale of training. The model was trained on a combination of several large text corpora, including Common Crawl, WebText2, Books1, Books2, and Wikipedia. The total data volume was approximately 570 GB or more (570 GB accounted for the filtered portion of Common Crawl, which dominated the training mixture).

One of the main features of GPT-3 was its capability for **few-shot learning** and **zero-shot learning**: the model could perform a wide range of natural language processing tasks, including translation, summarization, question answering, essay writing, and even programming, based on just a few examples in the text prompt or none at all.

## GPT-3.5

The **GPT-3.5** model was introduced by OpenAI in late 2022 as part of the evolutionary development of the GPT family. It was built on the scaled autoregressive decoder-only transformer architecture used in GPT-3, with improvements in text generation quality, context processing, and the ability to follow complex instructions. The exact number of parameters in GPT-3.5 was not officially disclosed; the davinci versions are presumably comparable in size to GPT-3 (175B), but the exact parameters of the gpt-3.5-turbo version are unknown.

The training of GPT-3.5 involved expanded use of Reinforcement Learning from Human Feedback (RLHF) methods in the text-davinci-003 and gpt-3.5-turbo versions. The earlier text-davinci-002 version was trained using supervised fine-tuning (SFT) rather than RLHF. The model was trained on expanded text corpora including Common Crawl, Books, WebText, and other high-quality sources. The context window in early popular versions (gpt-3.5-turbo) was 4,096 tokens; OpenAI subsequently released updated versions with a context of up to 16,385 tokens<sup>[\[10\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-10)</sup>.

In practice, GPT-3.5 was adapted to solve a wide range of natural language processing tasks, such as:

- Generating coherent and logical text;
- Question answering (QA) and context understanding;
- Following multi-step instructions;
- Improved long-term context maintenance in dialogues.

Several key versions based on GPT-3.5 were released for different purposes:

- **text-davinci-002** — the first publicly available model based on GPT-3.5, optimized for generation and instruction following (trained using SFT).
- **text-davinci-003** — an improved version with even greater reasoning and complex text generation capability (trained using RLHF).
- **gpt-3.5-turbo** — the most performant and cost-effective version of GPT-3.5, used in the ChatGPT service since late 2022.

## GPT-4

The **GPT-4** model was introduced by OpenAI on March 14, 2023 in the *"GPT-4 Technical Report"*. It represented the next stage in the development of the language model family, offering significant improvements in text comprehension, generation of meaningful and creative responses, and processing of multimodal data. The exact number of parameters and architectural details of the model were not officially disclosed — the GPT-4 technical report explicitly states that information about the architecture, model size, hardware, training compute, and dataset construction is not published<sup>[\[11\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-11)</sup>. According to unofficial external estimates, GPT-4 may have used a [Mixture of Experts](https://systems-analysis.info/eng/Mixture-of-Experts_(MoE) "Mixture-of-Experts (MoE)") (MoE) approach with a total scale on the order of ~1.8 trillion parameters; however, OpenAI has neither officially confirmed nor denied these figures<sup>[\[12\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-12)</sup>.

GPT-4 is a multimodal model capable of accepting both text and images as input. It should be noted that at the time of the initial launch in March 2023, only the text modality was available; image input support was deployed later. The context window was 8,192 tokens in the base version and 32,768 tokens in the GPT-4-32k variant. The model used RLHF (Reinforcement Learning from Human Feedback) methods.

GPT-4 training was performed on a combination of large-scale textual and multimodal corpora. Specific details of the training data, hardware, and methodology are not disclosed in official OpenAI publications.

Training proceeded in several stages:

- large-scale unsupervised pre-training on texts and images,
- supervised fine-tuning on specialized tasks,
- a final stage of Reinforcement Learning from Human Feedback (RLHF) to improve reliability, safety, and instruction interpretation quality.

Several main versions were released based on GPT-4:

- **GPT-4 (March 2023):** the base version with text input support (image support added later); context window of 8,192 tokens; a GPT-4-32k variant with a 32,768-token context was also released.
- **GPT-4 Turbo (November 2023):** an optimized modification of GPT-4 with an increased context window of up to 128,000 tokens<sup>[\[13\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-13)</sup>; reduced compute costs and accelerated generation; support for [function calling](https://systems-analysis.info/eng/Function_calling_(LLM) "Function calling (LLM)") and JSON output modes.
- **GPT-4o (May 2024):** a next-generation multimodal version; in the launch announcement it was positioned as an omni-model capable of working with text, images, and audio in real time (unlike GPT-4 Turbo, where different modalities were served by separate modules); however, the base API model `gpt-4o` is described as text+image input, text output; context window of 128,000 tokens.
- **GPT-4.5 (February 2025):** a research preview; the OpenAI system card explicitly states that the model "builds on GPT-4o"<sup>[\[3\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt45-systemcard-3)</sup>; improved generation of complex texts, increased instruction-following accuracy, and reduced hallucination rate; context window of 128,000 tokens. It was described as "the last OpenAI model without [chain-of-thought](https://systems-analysis.info/eng/Chain-of-Thought_Prompting "Chain-of-Thought Prompting")" (codename — *Orion*)<sup>[\[14\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-14)</sup>. The API model `gpt-4.5-preview` was declared deprecated on April 14, 2025 and shut down on July 14, 2025<sup>[\[4\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-deprecations-4)</sup>.
- **GPT-4.1 (April 2025):** a stable version with a radical context expansion to 1,047,576 tokens; accepts text and images as input, outputs text<sup>[\[15\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-15)</sup>; released simultaneously in three variants (GPT-4.1, GPT-4.1 mini, GPT-4.1 nano); initially available only through the API, later deployed in ChatGPT.

## GPT-5

**On August 7, 2025**, OpenAI introduced GPT‑5 as its then "smartest, fastest, and most useful" model, with a built-in deep reasoning mode (*thinking*) and a focus on practical scenarios — writing, programming, health-related work, and multimodal understanding. GPT‑5 gradually became the default model for most logged-in ChatGPT users, displacing the previously used GPT‑4/4o family and o-series models.<sup>[\[16\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-16)</sup>

GPT‑5 is implemented as a **unified system** with two main operating modes: fast, cost-effective responses for everyday queries (referred to as **gpt‑5 main**) and deep reasoning for complex tasks (referred to as **gpt‑5 thinking**). The mode is selected automatically by a router that takes into account the dialogue type, query complexity, the need for tools, and explicit user cues (e.g., "think step by step" or "analyze in depth"). In ChatGPT, users have access to Auto / Instant / Thinking / Pro modes; the **mini** and **nano** variants are primarily API models, and mini in the consumer product may be used as a fallback after rate limits are exhausted<sup>[\[17\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-chatgpt-53-54-17)</sup>.

Several sizes and configurations of GPT‑5 are available through the API; in the OpenAI documentation, the main variants are listed as **`gpt‑5`**, **`gpt‑5‑mini`**, and **`gpt‑5‑nano`** (all supporting text and visual data). The maximum total context window for the GPT‑5 family in the API is approximately **400,000 tokens** (with separate budgets for input and reasoning/output), although specific limits may vary depending on the chosen model variant and product<sup>[\[18\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-18)</sup>.

On a number of web-search and factual benchmarks, GPT‑5 demonstrates a notable reduction in the frequency of hallucinations and errors compared to GPT‑4o and earlier OpenAI "thinking" models. In the official announcement, OpenAI reported error reductions of approximately 45% compared to GPT-4o and approximately 80% compared to o3 in thinking mode — these results were obtained under specific conditions: with web search enabled on anonymized prompts representative of ChatGPT production traffic<sup>[\[19\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-19)</sup>.

## GPT-5.1

The **GPT-5.1** model was introduced by OpenAI on November 12, 2025 as the first significant iteration after the base GPT-5, aimed at improving everyday interaction, conversational quality, and adaptability. The model retains the unified system with a fast mode (**GPT-5.1 Instant**) and deep reasoning (**GPT-5.1 Thinking**), but introduces **adaptive reasoning**: the model dynamically determines the amount of computation depending on query complexity, making it notably faster on simple tasks without sacrificing quality on complex ones.

GPT-5.1 training was built on top of GPT-5 with an additional post-training stage that included expanded RLHF, a focus on natural tone, and reduced "coldness" of responses. The API context window is 400,000 tokens, with a maximum output of 128,000 tokens<sup>[\[20\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt51-api-20)</sup>. Extended prompt caching of up to 24 hours was introduced, significantly reducing cost and latency for multi-turn dialogues<sup>[\[21\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt51-announce-21)</sup>.

Key features:

- **GPT-5.1 Instant** — the primary mode for everyday tasks; the first to use adaptive reasoning to determine when it is worth "thinking" before responding to a more complex query<sup>[\[21\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt51-announce-21)</sup>.
- **GPT-5.1 Thinking** — adaptive allocation of reasoning time; according to OpenAI, on a representative distribution of ChatGPT tasks the model is approximately twice as fast on the simplest tasks and approximately twice as slow on the hardest compared to GPT-5 Thinking<sup>[\[21\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt51-announce-21)</sup>.
- Improved multimodality (text + vision).
- Improved coding and agentic scenarios, as well as efficiency on simple tasks through adaptive reasoning and extended prompt caching<sup>[\[22\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt51-dev-22)</sup>.

## GPT-5.2

The **GPT-5.2** model was released on December 11, 2025 as "the most capable model in the series for professional work and learning." It is an evolution of GPT-5.1 with an emphasis on economic value: generation of tables, presentations, complex code, and end-to-end tasks. It retains the unified architecture with **Instant**, **Thinking**, and a new **Pro** mode (for tasks requiring maximum compute and reasoning time).

Training included an updated corpus with a knowledge cutoff of August 2025, enhanced instruction-tuning, and RLHF to reduce errors in multi-step scenarios. Context window — 400K tokens (128K max output). The model became more reliable in professional scenarios, with improved factual accuracy and tool use.

On December 18, 2025, the specialized **GPT-5.2-Codex** was released based on GPT-5.2 — an agentic coding model with improved context compaction, Windows support, enhanced cybersecurity, and long-horizon reasoning (tasks lasting up to several hours).

As of February 13, 2026, following the retirement of several older models, GPT-5.2 temporarily became the default model in ChatGPT. However, by early March 2026, this role was taken over by GPT‑5.3 Instant and GPT‑5.4<sup>[\[17\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-chatgpt-53-54-17)</sup>.

## GPT-5.3-Codex

**GPT-5.3-Codex** was introduced on February 5, 2026 as "the most powerful agentic coding model to date." It combines the frontier-coding capabilities of GPT-5.2-Codex with the professional reasoning of GPT-5.2 in a single model that is 25% faster than its predecessors.

The model is capable of performing virtually any developer task: long-running workflows, research, tool use, code execution, and interactive steering (the user can intervene in real time without losing context). Early versions of the model were used by the OpenAI team to debug their own training, deployment, and evaluations.

Key results at the time of the February 5, 2026 announcement: Terminal-Bench ~77.3%, OSWorld-Verified ~64.7%, SWE-Bench Pro ~56.8%. In the later GPT-5.4 release on March 5, 2026, OpenAI reported an updated OSWorld-Verified result of 74.0% for GPT-5.3-Codex when using a new API parameter that preserves the original image resolution<sup>[\[23\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-23)[\[7\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt54-announce-7)</sup>.

On February 12, 2026, OpenAI also released **GPT-5.3-Codex-Spark** — a compact ultra-fast version in partnership with Cerebras, optimized for real-time use: over 1,000 tokens per second, text-only, 128K context. At launch, this was a rollout for ChatGPT Pro users in Codex and a small number of API design partners, rather than a broadly available API model<sup>[\[24\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-24)</sup>.

## GPT-5.4

**On March 5, 2026**, OpenAI introduced GPT‑5.4 as its new frontier model for professional work. GPT‑5.4 combines the strengths of OpenAI's latest releases in reasoning, coding, and agentic workflows and was the first in the main product line to receive built-in computer use capabilities. Simultaneously, OpenAI released **GPT‑5.4 Pro** — a variant for the most complex tasks, using more compute and longer reasoning<sup>[\[7\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt54-announce-7)</sup>.

In the API, the **`gpt-5.4`** model is described as the recommended default for a wide range of general-purpose and coding tasks; the context window is 1,050,000 tokens, with a maximum output of 128,000 tokens. The model accepts text and images as input and outputs text<sup>[\[8\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt54-api-8)[\[25\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt54-model-25)</sup>.

In ChatGPT, the **Auto** mode as of March 7, 2026 automatically switches between **GPT‑5.3 Instant** and **GPT‑5.4 Thinking**, while **GPT‑5.4 Pro** is available as a separate high-capability mode. For logged-in ChatGPT users, the default model is GPT‑5.3<sup>[\[17\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-chatgpt-53-54-17)</sup>.

## GPT-5.5

**On April 23, 2026**, OpenAI introduced GPT-5.5 as its new frontier model for professional work and described it as a new class of intelligence for real work. GPT-5.5 was positioned as a step forward in agentic coding, computer use, knowledge work, and early scientific research, with stronger ability to plan, use tools, check its work, and carry out multi-step tasks across software and documents. OpenAI also released **GPT-5.5 Pro** — a higher-compute variant for harder questions and higher-accuracy work. On April 24, 2026, both GPT-5.5 and GPT-5.5 Pro became available in the API<sup>[\[26\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt55-announce-26)</sup>.

In the API, the **`gpt-5.5`** model is described as OpenAI's newest frontier model for the most complex professional work. The context window is 1,050,000 tokens, with a maximum output of 128,000 tokens. The model accepts text and images as input and outputs text; `reasoning.effort` supports `none`, `low`, `medium` by default, `high`, and `xhigh`<sup>[\[27\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt55-api-27)</sup>. The **`gpt-5.5-pro`** model uses more compute to think harder and provide more consistently precise answers; it is available for Responses API requests, including through the Batch API, and has the same 1,050,000-token context window and 128,000-token maximum output<sup>[\[28\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt55-pro-api-28)</sup>.

In ChatGPT, **GPT-5.5 Thinking** was made available to Plus, Pro, Business, and Enterprise users, while **GPT-5.5 Pro** was made available to Pro, Business, and Enterprise users. In Codex, GPT-5.5 was made available to Plus, Pro, Business, Enterprise, Edu, and Go plans with a 400K context window<sup>[\[26\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt55-announce-26)</sup>. On May 5, 2026, OpenAI introduced **GPT-5.5 Instant** as the updated default ChatGPT model, replacing **GPT-5.3 Instant**; in the API, the corresponding latest Instant model is exposed through **`chat-latest`**<sup>[\[29\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt55-instant-29)</sup>.

## GPT-5.6

**On June 26, 2026**, OpenAI began a limited preview of the GPT-5.6 family, comprising **GPT-5.6 Sol**, **GPT-5.6 Terra**, and **GPT-5.6 Luna**. On **July 9, 2026**, the family entered general availability across ChatGPT, Codex, and the OpenAI API. Sol was introduced as the flagship model for complex professional work, Terra as a balanced lower-cost model for everyday workloads, and Luna as the fastest and most cost-efficient member of the family. OpenAI positioned GPT-5.6 as an advance in coding, long-horizon planning, computer use, professional knowledge work, design, cybersecurity, and scientific research, with improved performance per token and stronger ability to complete multi-step tasks using tools<sup>[\[30\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt56-announce-30)</sup>.

GPT-5.6 introduced **`max`** as a reasoning-effort setting above `xhigh`, allowing the models to spend additional time exploring alternatives, checking results, and revising their approach. OpenAI also introduced **`ultra`**, a higher-capability multi-agent mode that coordinates four agents in parallel by default. In the Responses API, **Programmatic Tool Calling** allows GPT-5.6 to write and execute lightweight in-memory programs that coordinate tools and process intermediate results, while a beta multi-agent capability allows the model to run concurrent subagents and synthesize their work into a single response. OpenAI stated that GPT-5.6 was released with its most extensive safety evaluation and safeguard system to date<sup>[\[30\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt56-announce-30)</sup>.

In the API, **`gpt-5.6-sol`** is the frontier model for complex professional work, and the **`gpt-5.6`** alias routes requests to Sol. **`gpt-5.6-terra`** is intended to balance intelligence and cost and broadly occupies the position of the `mini` tier used in earlier GPT-5 families, while **`gpt-5.6-luna`** is optimized for cost-sensitive, high-volume workloads and broadly corresponds to the earlier `nano` tier. All three models have a 1,050,000-token context window, a maximum output of 128,000 tokens, and a knowledge cutoff of February 16, 2026. They accept text and images as input, output text, and support `reasoning.effort` values of `none`, `low`, `medium`, `high`, `xhigh`, and `max`<sup>[\[31\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt56-sol-api-31)[\[32\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt56-terra-api-32)[\[33\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt56-luna-api-33)</sup>.

In ChatGPT, Plus, Pro, Business, and Enterprise users received access to GPT-5.6 Sol through medium and higher reasoning-effort settings, while Pro and Enterprise users could also select **GPT-5.6 Sol Pro** for higher-quality results on difficult tasks. Sol Pro is a higher-compute product mode rather than a separately documented API model. In ChatGPT Work and Codex, Free and Go users received access to GPT-5.6 Terra, while Plus, Pro, Business, and Enterprise users could choose between Sol, Terra, and Luna. The `max` setting was made available to users with GPT-5.6 access in ChatGPT Work and Codex; `ultra` was made available to Pro and Enterprise users in ChatGPT Work and to Plus and higher plans in Codex<sup>[\[30\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt56-announce-30)</sup>.

## GPT-6

**On September 3, 2026**, OpenAI introduced **GPT-6 Astra**, the first publicly announced model of the GPT-6 generation, and began a limited rollout to selected organizations. OpenAI described Astra as its most capable and most aligned model to date and positioned it as the new flagship for the hardest end-to-end work. The model was introduced with advances in computer use, browsing, software engineering, coding, cybersecurity, scientific research, and professional knowledge work, with particular emphasis on completing complex multi-step workflows across browsers, code, professional software, and document-creation tools. Astra can create and edit documents, spreadsheets, presentations, websites, and other professional artifacts while preserving templates and incorporating new instructions without losing track of the broader task<sup>[\[34\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt6-announce-34)</sup>.

GPT-6 Astra introduced several capabilities intended for longer-running agentic workflows. **Async tool calling** allows the model to continue reasoning, invoke other tools, or work on independent parts of a task while an application is executing a tool call. **Mid-turn steering** allows additional user instructions to be incorporated while a response is already in progress, preserving completed work rather than restarting the task. A new **`configuration_update`** mechanism allows reasoning effort to be changed during a conversation while preserving the existing prompt cache. Astra retains capabilities introduced with GPT-5.6, including **Programmatic Tool Calling**, multi-agent orchestration, computer use, Structured Outputs, streaming, persisted reasoning, prompt caching, compaction, and pro mode. Unlike GPT-5.6, Astra does not support the `none` reasoning-effort setting; supported `reasoning.effort` values are `low`, `medium`, `high`, `xhigh`, and `max`<sup>[\[35\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt6-guide-35)</sup>.

In the API, **`gpt-6-astra`** is the flagship GPT-6 model for complex reasoning, coding, computer use, research, and document creation. It has a **1,050,000-token context window**, a maximum output of **128,000 tokens**, and a knowledge cutoff of **April 30, 2026**. Astra accepts text and images as input and outputs text, and supports streaming, function calling, Structured Outputs, reasoning tokens, and tools including web search, file search, and computer use. Standard API pricing at launch was **\$10 per million input tokens**, **\$1 per million cached input tokens**, and **\$50 per million output tokens**; prompts exceeding 272,000 input tokens use higher long-context rates. Astra is available through the Responses API and Chat Completions, although OpenAI recommends the Responses API for tool-using applications<sup>[\[36\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt6-api-36)[\[35\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt6-guide-35)</sup>.

At launch, GPT-6 Astra began rolling out to a limited set of organizations, with broader access announced for **ChatGPT Plus, Pro, Business, and Enterprise** users and through the **OpenAI API, Microsoft Azure, and AWS Bedrock** over the following days. Users on Pro, Business, and Enterprise plans were also announced to receive a higher-compute **GPT-6 Astra Pro** mode; OpenAI's subsequent ChatGPT documentation refers to this product mode as **GPT-6 Pro, powered by GPT-6 Astra**. Plus users receive Astra in ChatGPT Work and Codex as the rollout reaches their accounts, while availability can differ between Chat, Work, and Codex during the staged deployment. Enterprise administrators control whether Astra is enabled for their workspace<sup>[\[34\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt6-announce-34)[\[37\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt6-chatgpt-37)</sup>.

GPT-6 Astra was also the first OpenAI model classified at the **Critical** level for cybersecurity capability under the company's Preparedness Framework. OpenAI reported that, with appropriate tools and access, the model can identify previously unknown vulnerabilities and develop new exploitation techniques across hardened systems. The deployment therefore introduced strengthened cybersecurity safeguards and **misalignment monitoring**, including monitoring of agent trajectories and additional checks that can pause or stop tasks when potentially unauthorized behavior is detected. OpenAI stated that Astra showed substantially improved adherence to authorized task boundaries compared with GPT-5.6 Sol in its internal alignment evaluations<sup>[\[38\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt6-safety-38)</sup>.

  

## GPT model evolution

The table covers the principal general-purpose GPT language-model lineage and its officially named size, reasoning and product variants. Voice, image-generation, embedding, moderation and search-only models are outside its scope.

| Model / variant            | Initial release  | Parameter count             | Publicly disclosed training data or knowledge cutoff                                                    | Context window                                              | Key features and status                                                                                                                                                                                                                       |
|----------------------------|------------------|-----------------------------|---------------------------------------------------------------------------------------------------------|-------------------------------------------------------------|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| GPT-1                      | June 2018        | ≈117M                       | BooksCorpus: more than 7,000 unpublished books                                                          | 512 tokens                                                  | Decoder-only Transformer; generative pre-training followed by supervised task-specific fine-tuning                                                                                                                                            |
| GPT-2                      | 14 February 2019 | 1.5B in the largest version | WebText: approximately 40 GB from about 8 million web pages                                             | 1,024 tokens                                                | Substantially improved coherent text generation and zero-shot behavior; initially released in stages because of misuse concerns                                                                                                               |
| GPT-3                      | 28 May 2020      | 175B                        | Common Crawl, WebText2, Books1, Books2 and Wikipedia; approximately 300B tokens sampled during training | 2,048 tokens                                                | Large-scale in-context learning; strong few-shot and zero-shot capabilities without gradient-based task fine-tuning                                                                                                                           |
| GPT-3.5                    | 2022             | Not disclosed               | Not disclosed                                                                                           | Approximately 4K–16K tokens, depending on the model version | Instruction tuning and reinforcement learning from human feedback; improved dialogue and instruction following; foundation of the original ChatGPT                                                                                            |
| GPT-3.5 Turbo              | 1 March 2023     | Not disclosed               | Not disclosed                                                                                           | Initially 4,096 tokens; later versions up to 16,385 tokens  | Lower-cost chat-optimized GPT-3.5 model; later made available for fine-tuning; subsequently deprecated                                                                                                                                        |
| GPT-4                      | 14 March 2023    | Not disclosed               | Not disclosed                                                                                           | 8,192 or 32,768 tokens                                      | Multimodal text-and-image input; improved reasoning, reliability and instruction following                                                                                                                                                    |
| GPT-4 Turbo Preview        | 6 November 2023  | Not disclosed               | Knowledge through April 2023 at launch                                                                  | 128,000 tokens                                              | Preview of the faster and less expensive GPT-4 Turbo family; substantially enlarged context window                                                                                                                                            |
| GPT-4 Turbo                | 9 April 2024     | Not disclosed               | Knowledge cutoff: December 2023 for the final API snapshot                                              | 128,000 tokens                                              | Production GPT-4 Turbo model with text and image input; lower cost than the original GPT-4                                                                                                                                                    |
| GPT-4o                     | 13 May 2024      | Not disclosed               | Knowledge cutoff: 1 October 2023 for the principal API snapshot                                         | 128,000 tokens                                              | End-to-end multimodal model spanning text, vision and audio; lower latency and improved real-time interaction                                                                                                                                 |
| GPT-4o mini                | 18 July 2024     | Not disclosed               | Knowledge cutoff: 1 October 2023                                                                        | 128,000 tokens                                              | Smaller and less expensive GPT-4o-family model intended for focused and high-volume tasks                                                                                                                                                     |
| GPT-4o Search Preview      | March 2025       | Not disclosed               | Knowledge cutoff: 1 October 2023                                                                        | 128,000 tokens                                              | Specialized GPT-4o variant trained to formulate and execute web searches through the Chat Completions API; subsequently deprecated                                                                                                            |
| GPT-4o mini Search Preview | March 2025       | Not disclosed               | Knowledge cutoff: 1 October 2023                                                                        | 128,000 tokens                                              | Lower-cost search-specialized GPT-4o mini variant; subsequently deprecated                                                                                                                                                                    |
| GPT-4.5 Preview            | 27 February 2025 | Not disclosed               | Knowledge cutoff: 1 October 2023                                                                        | 128,000 tokens                                              | Research preview emphasizing scaled unsupervised learning, broader knowledge, natural conversation and reduced hallucinations; later discontinued                                                                                             |
| GPT-4.1                    | 14 April 2025    | Not disclosed               | Knowledge cutoff: 1 June 2024                                                                           | 1,047,576 tokens                                            | Major improvements in coding, instruction following and tool calling; text and image input with text output                                                                                                                                   |
| GPT-4.1 mini               | 14 April 2025    | Not disclosed               | Knowledge cutoff: 1 June 2024                                                                           | 1,047,576 tokens                                            | Faster and less expensive GPT-4.1 variant retaining the approximately one-million-token context window                                                                                                                                        |
| GPT-4.1 nano               | 14 April 2025    | Not disclosed               | Knowledge cutoff: 1 June 2024                                                                           | 1,047,576 tokens                                            | Smallest and fastest GPT-4.1 variant, intended for classification, extraction, ranking and other high-volume tasks                                                                                                                            |
| GPT-5 Instant              | 7 August 2025    | Not disclosed               | Knowledge cutoff: 30 September 2024                                                                     | 128,000 tokens in the corresponding Chat API model          | Fast conversational component of the original unified GPT-5 system; handled ordinary queries and could route harder requests to GPT-5 Thinking                                                                                                |
| GPT-5 Thinking             | 7 August 2025    | Not disclosed               | Knowledge cutoff: 30 September 2024                                                                     | 400,000 tokens                                              | Reasoning component of the unified GPT-5 system; configurable reasoning effort, improved coding, tool use and factual reliability                                                                                                             |
| GPT-5 mini                 | 7 August 2025    | Not disclosed               | Knowledge cutoff: 31 May 2024                                                                           | 400,000 tokens                                              | Faster and more cost-efficient GPT-5 variant for well-defined, latency-sensitive and high-volume workloads                                                                                                                                    |
| GPT-5 nano                 | 7 August 2025    | Not disclosed               | Knowledge cutoff: 31 May 2024                                                                           | 400,000 tokens                                              | Fastest and least expensive original GPT-5 variant; optimized for summarization, classification and extraction                                                                                                                                |
| GPT-5 Pro                  | 2025             | Not disclosed               | Knowledge cutoff: 30 September 2024                                                                     | 400,000 tokens                                              | Higher-compute GPT-5 variant for difficult reasoning tasks; API version supports only high reasoning effort                                                                                                                                   |
| GPT-5.1 Instant            | 12 November 2025 | Not disclosed               | Knowledge cutoff: 30 September 2024                                                                     | 128,000 tokens in the corresponding Chat API model          | More conversational fast model with improved instruction following and adaptive reasoning                                                                                                                                                     |
| GPT-5.1 Thinking           | 12 November 2025 | Not disclosed               | Knowledge cutoff: 30 September 2024                                                                     | 400,000 tokens                                              | Dynamically adjusted reasoning time according to task difficulty; clearer explanations, improved coding and reduced unnecessary reasoning                                                                                                     |
| GPT-5.1 Pro                | November 2025    | Not disclosed               | Not disclosed separately                                                                                | Approximately 400,000 tokens                                | Higher-compute GPT-5.1 variant for difficult questions; retired from ChatGPT on 11 March 2026                                                                                                                                                 |
| GPT-5.2 Instant            | 11 December 2025 | Not disclosed               | Knowledge cutoff: 31 August 2025                                                                        | 128,000 tokens in the corresponding Chat API model          | Fast conversational GPT-5.2 variant; improved everyday usefulness, information retrieval and instruction following                                                                                                                            |
| GPT-5.2 Thinking           | 11 December 2025 | Not disclosed               | Knowledge cutoff: 31 August 2025                                                                        | 400,000 tokens                                              | Professional knowledge work, coding, long-context reasoning, vision and tool use; introduced the *xhigh* reasoning-effort setting                                                                                                             |
| GPT-5.2 Pro                | 11 December 2025 | Not disclosed               | Knowledge cutoff: 31 August 2025                                                                        | 400,000 tokens                                              | Higher-compute GPT-5.2 variant for complex questions where accuracy is prioritized over latency                                                                                                                                               |
| GPT-5.3 Instant            | 3 March 2026     | Not disclosed               | Knowledge cutoff: 31 August 2025                                                                        | 128,000 tokens                                              | Faster conversational model with improved factuality, web-search integration and conversational flow; later superseded by GPT-5.5 Instant                                                                                                     |
| GPT-5.4                    | 5 March 2026     | Not disclosed               | Knowledge cutoff: 31 August 2025                                                                        | 1,050,000 tokens                                            | Frontier model for professional and agentic work; native computer use, stronger coding, multimodal reasoning and coordinated tool execution                                                                                                   |
| GPT-5.4 Pro                | 5 March 2026     | Not disclosed               | Knowledge cutoff: 31 August 2025                                                                        | 1,050,000 tokens                                            | Higher-compute GPT-5.4 variant producing more precise responses on difficult tasks                                                                                                                                                            |
| GPT-5.4 mini               | 17 March 2026    | Not disclosed               | Knowledge cutoff: 31 August 2025                                                                        | 400,000 tokens                                              | Faster model for coding, computer use, multimodal workloads and sub-agent systems; supports text and image input and major API tools                                                                                                          |
| GPT-5.4 nano               | 17 March 2026    | Not disclosed               | Knowledge cutoff: 31 August 2025                                                                        | 400,000 tokens                                              | Lowest-cost GPT-5.4-class model for classification, extraction, ranking and high-volume sub-agent tasks                                                                                                                                       |
| GPT-5.5                    | 23 April 2026    | Not disclosed               | Knowledge cutoff: 1 December 2025                                                                       | 1,050,000 tokens                                            | Frontier model for complex professional work, agentic coding, research, data analysis, document creation, computer use and multi-tool workflows                                                                                               |
| GPT-5.5 Pro                | 23 April 2026    | Not disclosed               | Knowledge cutoff: 1 December 2025                                                                       | 1,050,000 tokens                                            | Higher-compute GPT-5.5 variant designed for increased accuracy on difficult and long-running tasks                                                                                                                                            |
| GPT-5.5 Instant            | 5 May 2026       | Not disclosed               | Knowledge cutoff: 31 August 2025 for the corresponding dynamic API alias                                | 400,000 tokens                                              | Default conversational ChatGPT model after launch; improved factuality, concision, personalization, image understanding, STEM answers and web-search decisions                                                                                |
| GPT-5.6 Sol                | 9 July 2026      | Not disclosed               | Knowledge cutoff: 16 February 2026                                                                      | 1,050,000 tokens                                            | Flagship GPT-5.6 tier for complex professional work; stronger reasoning, coding, science, cybersecurity, computer use, design judgment and token efficiency                                                                                   |
| GPT-5.6 Sol Pro            | 9 July 2026      | Not disclosed               | Knowledge cutoff: 16 February 2026                                                                      | 1,050,000 tokens                                            | Higher-compute product mode for the most difficult tasks; based on GPT-5.6 Sol rather than a separately disclosed base architecture                                                                                                           |
| GPT-5.6 Terra              | 9 July 2026      | Not disclosed               | Knowledge cutoff: 16 February 2026                                                                      | 1,050,000 tokens                                            | Balanced GPT-5.6 tier combining high capability with lower cost; broadly corresponds to the mini position in earlier GPT-5 families                                                                                                           |
| GPT-5.6 Luna               | 9 July 2026      | Not disclosed               | Knowledge cutoff: 16 February 2026                                                                      | 1,050,000 tokens                                            | Fastest and most cost-efficient GPT-5.6 tier for cost-sensitive, high-volume workloads; broadly corresponds to the nano position in earlier GPT-5 families                                                                                    |
| GPT-6 Astra                | 3 September 2026 | Not disclosed               | Knowledge cutoff: 30 April 2026                                                                         | 1,050,000 tokens                                            | Flagship GPT-6 model for the hardest end-to-end work; state-of-the-art computer use, browsing, software engineering, cybersecurity, science and professional work; supports long-horizon multi-step workflows and up to 128,000 output tokens |
| GPT-6 Pro                  | 3 September 2026 | Not disclosed               | Knowledge cutoff: 30 April 2026 for the underlying GPT-6 Astra model                                    | 1,050,000 tokens for the underlying GPT-6 Astra model       | Higher-compute ChatGPT product mode powered by GPT-6 Astra for the most difficult tasks and longer-running workflows; rolling out to eligible Pro, Business and Enterprise plans                                                              |

GPT model evolution

### Specialized GPT and Codex models

| Model               | Initial release   | Parameter count | Context window                     | Distinguishing role                                                                                                                                                                                                                                 |
|---------------------|-------------------|-----------------|------------------------------------|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| GPT-5-Codex         | 15 September 2025 | Not disclosed   | 400,000 tokens                     | GPT-5 variant optimized through reinforcement learning for agentic software-engineering tasks, code review, terminal use and iterative testing                                                                                                      |
| GPT-5.1-Codex       | 13 November 2025  | Not disclosed   | 400,000 tokens                     | GPT-5.1 variant optimized for long-running agentic coding tasks in Codex and similar harnesses                                                                                                                                                      |
| GPT-5.1-Codex mini  | 13 November 2025  | Not disclosed   | 400,000 tokens                     | Smaller and more cost-efficient GPT-5.1-Codex variant for less complex coding workflows                                                                                                                                                             |
| GPT-5.1-Codex-Max   | 19 November 2025  | Not disclosed   | 400,000 tokens per context segment | Frontier agentic coding model for project-scale refactors, deep debugging and tasks lasting many hours; supports automatic context compaction                                                                                                       |
| GPT-5.2-Codex       | 18 December 2025  | Not disclosed   | 400,000 tokens                     | Long-horizon agentic coding, large refactors, migrations, Windows development, visual understanding and defensive-cybersecurity capabilities                                                                                                        |
| GPT-5.3-Codex       | 5 February 2026   | Not disclosed   | 400,000 tokens                     | Combined agentic coding with broader reasoning and professional knowledge; supported interactive steering during long-running work                                                                                                                  |
| GPT-5.3-Codex-Spark | 12 February 2026  | Not disclosed   | 128,000 tokens                     | Compact text-only real-time coding model designed for interactive steering and very high generation speed                                                                                                                                           |
| GPT-5.5-Cyber       | 2026              | Not disclosed   | Not publicly specified             | Specialized GPT-5.5 model for vetted professionals performing advanced authorized cybersecurity work                                                                                                                                                |
| GPT-5.6 Cyber       | 10 August 2026    | Not disclosed   | 400,000 tokens                     | Purpose-trained GPT-5.6 Sol-based cybersecurity model for approved defenders conducting advanced authorized vulnerability research, exploit validation and security testing; available through Daybreak Red with separate approval and provisioning |

Specialized GPT and Codex models

## Architectural parameters of GPT models

### GPT models with published architectural parameters

| Model | Release year | Parameter count | Number of layers | Hidden state size | Number of attention heads | Context window | Training corpus size                                       |
|-------|--------------|-----------------|------------------|-------------------|---------------------------|----------------|------------------------------------------------------------|
| GPT-1 | 2018         | ≈117M           | 12               | 768               | 12                        | 512 tokens     | BooksCorpus; over 7,000 unpublished books                  |
| GPT-2 | 2019         | 1.5B / 1.558B   | 48               | 1,600             | 25                        | 1,024 tokens   | WebText; slightly over 8 million documents / 40 GB of text |
| GPT-3 | 2020         | 175B            | 96               | 12,288            | 96                        | 2,048 tokens   | ≈499B-token source mixture; trained for 300B tokens        |

GPT models with published architectural parameters

<table class="wikitable" style="text-align:center; width:100%;">
<caption>GPT models without disclosed architectural parameters (as of 2026-07-10)</caption>
<colgroup>
<col style="width: 9%" />
<col style="width: 9%" />
<col style="width: 9%" />
<col style="width: 9%" />
<col style="width: 9%" />
<col style="width: 9%" />
<col style="width: 9%" />
<col style="width: 9%" />
<col style="width: 9%" />
<col style="width: 9%" />
<col style="width: 9%" />
</colgroup>
<thead>
<tr class="header">
<th>Model</th>
<th>Release year</th>
<th>Public model ID / alias</th>
<th>API status</th>
<th>Context window</th>
<th>Max output</th>
<th>Input modalities</th>
<th>Output modalities</th>
<th>Knowledge cutoff</th>
<th>Reasoning effort</th>
<th>API doc</th>
</tr>
</thead>
<tbody>
<tr class="odd">
<td>GPT-3.5 Turbo</td>
<td>2023</td>
<td><code>gpt-3.5-turbo</code></td>
<td><a href="https://developers.openai.com/api/docs/deprecations" class="external text" rel="nofollow">Deprecated</a><br />
shutdown 2026-10-23</td>
<td>16,385 tokens</td>
<td>4,096 tokens</td>
<td>Text</td>
<td>Text</td>
<td>2021-09-01</td>
<td>—</td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-3.5-turbo" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="even">
<td>GPT-4</td>
<td>2023</td>
<td><code>gpt-4</code></td>
<td><a href="https://developers.openai.com/api/docs/deprecations" class="external text" rel="nofollow">Deprecated</a><br />
shutdown 2026-10-23</td>
<td>8,192 tokens</td>
<td>8,192 tokens</td>
<td>Text</td>
<td>Text</td>
<td>2023-12-01</td>
<td>—</td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-4" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="odd">
<td>GPT-4 Turbo</td>
<td>2023</td>
<td><code>gpt-4-turbo</code></td>
<td><a href="https://developers.openai.com/api/docs/deprecations" class="external text" rel="nofollow">Deprecated</a><br />
shutdown 2026-10-23</td>
<td>128,000 tokens</td>
<td>4,096 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2023-12-01</td>
<td>—</td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-4-turbo" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="even">
<td>GPT-4o</td>
<td>2024</td>
<td><code>gpt-4o</code></td>
<td><a href="https://developers.openai.com/api/docs/models/all" class="external text" rel="nofollow">Deprecated</a></td>
<td>128,000 tokens</td>
<td>16,384 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2023-10-01</td>
<td>—</td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-4o" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="odd">
<td>GPT-4o mini</td>
<td>2024</td>
<td><code>gpt-4o-mini</code></td>
<td>Available</td>
<td>128,000 tokens</td>
<td>16,384 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2023-10-01</td>
<td>—</td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-4o-mini" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="even">
<td>GPT-4.5 Preview</td>
<td>2025</td>
<td><code>gpt-4.5-preview</code></td>
<td><a href="https://developers.openai.com/api/docs/deprecations" class="external text" rel="nofollow">Removed</a><br />
2025-07-14</td>
<td>128,000 tokens</td>
<td>16,384 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2023-10-01</td>
<td>—</td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-4.5-preview" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="odd">
<td>GPT-4.1</td>
<td>2025</td>
<td><code>gpt-4.1</code></td>
<td>Available</td>
<td>1,047,576 tokens</td>
<td>32,768 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2024-06-01</td>
<td>—</td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-4.1" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="even">
<td>GPT-4.1 mini</td>
<td>2025</td>
<td><code>gpt-4.1-mini</code></td>
<td>Available</td>
<td>1,047,576 tokens</td>
<td>32,768 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2024-06-01</td>
<td>—</td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-4.1-mini" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="odd">
<td>GPT-4.1 nano</td>
<td>2025</td>
<td><code>gpt-4.1-nano</code></td>
<td><a href="https://developers.openai.com/api/docs/deprecations" class="external text" rel="nofollow">Deprecated</a><br />
shutdown 2026-10-23</td>
<td>1,047,576 tokens</td>
<td>32,768 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2024-06-01</td>
<td>—</td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-4.1-nano" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="even">
<td>GPT-5</td>
<td>2025</td>
<td><code>gpt-5</code></td>
<td>Available</td>
<td>400,000 tokens</td>
<td>128,000 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2024-09-30</td>
<td><code>minimal</code>, <code>low</code>, <code>medium</code> (default), <code>high</code></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="odd">
<td>GPT-5 mini</td>
<td>2025</td>
<td><code>gpt-5-mini</code></td>
<td>Available</td>
<td>400,000 tokens</td>
<td>128,000 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2024-05-31</td>
<td><code>minimal</code>, <code>low</code>, <code>medium</code> (default), <code>high</code></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5-mini" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="even">
<td>GPT-5 nano</td>
<td>2025</td>
<td><code>gpt-5-nano</code></td>
<td>Available</td>
<td>400,000 tokens</td>
<td>128,000 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2024-05-31</td>
<td><code>minimal</code>, <code>low</code>, <code>medium</code> (default), <code>high</code></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5-nano" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="odd">
<td>GPT-5 Pro</td>
<td>2025</td>
<td><code>gpt-5-pro</code></td>
<td>Available</td>
<td>400,000 tokens</td>
<td>272,000 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2024-09-30</td>
<td><code>high</code> only (default)</td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5-pro" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="even">
<td>GPT-5 Chat</td>
<td>2025</td>
<td><code>gpt-5-chat-latest</code></td>
<td><a href="https://developers.openai.com/api/docs/deprecations" class="external text" rel="nofollow">Deprecated</a><br />
shutdown 2026-07-23</td>
<td>128,000 tokens</td>
<td>16,384 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2024-09-30</td>
<td>—</td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5-chat-latest" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="odd">
<td>GPT-5.1</td>
<td>2025</td>
<td><code>gpt-5.1</code></td>
<td>Available</td>
<td>400,000 tokens</td>
<td>128,000 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2024-09-30</td>
<td><code>none</code> (default), <code>low</code>, <code>medium</code>, <code>high</code></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.1" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="even">
<td>GPT-5.1 Chat</td>
<td>2025</td>
<td><code>gpt-5.1-chat-latest</code></td>
<td><a href="https://developers.openai.com/api/docs/deprecations" class="external text" rel="nofollow">Deprecated</a><br />
shutdown 2026-07-23</td>
<td>128,000 tokens</td>
<td>16,384 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2024-09-30</td>
<td>—</td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.1-chat-latest" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="odd">
<td>GPT-5.2</td>
<td>2025</td>
<td><code>gpt-5.2</code></td>
<td>Available</td>
<td>400,000 tokens</td>
<td>128,000 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2025-08-31</td>
<td><code>none</code> (default), <code>low</code>, <code>medium</code>, <code>high</code>, <code>xhigh</code></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.2" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="even">
<td>GPT-5.2 Chat</td>
<td>2025</td>
<td><code>gpt-5.2-chat-latest</code></td>
<td><a href="https://developers.openai.com/api/docs/deprecations" class="external text" rel="nofollow">Deprecated</a><br />
shutdown 2026-08-10</td>
<td>128,000 tokens</td>
<td>16,384 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2025-08-31</td>
<td>—</td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.2-chat-latest" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="odd">
<td>GPT-5.2 Pro</td>
<td>2025</td>
<td><code>gpt-5.2-pro</code></td>
<td>Available</td>
<td>400,000 tokens</td>
<td>128,000 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2025-08-31</td>
<td><code>medium</code>, <code>high</code>, <code>xhigh</code></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.2-pro" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="even">
<td>GPT-5.3 Chat / GPT-5.3 Instant</td>
<td>2026</td>
<td><code>gpt-5.3-chat-latest</code></td>
<td><a href="https://developers.openai.com/api/docs/deprecations" class="external text" rel="nofollow">Deprecated</a><br />
shutdown 2026-08-10</td>
<td>128,000 tokens</td>
<td>16,384 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2025-08-31</td>
<td>—</td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.3-chat-latest" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="odd">
<td>GPT-5.3-Codex</td>
<td>2026</td>
<td><code>gpt-5.3-codex</code></td>
<td>Available</td>
<td>400,000 tokens</td>
<td>128,000 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2025-08-31</td>
<td><code>low</code>, <code>medium</code>, <code>high</code>, <code>xhigh</code></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.3-codex" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="even">
<td>GPT-5.4</td>
<td>2026</td>
<td><code>gpt-5.4</code></td>
<td>Available</td>
<td>1,050,000 tokens</td>
<td>128,000 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2025-08-31</td>
<td><code>none</code> (default), <code>low</code>, <code>medium</code>, <code>high</code>, <code>xhigh</code></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.4" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="odd">
<td>GPT-5.4 Pro</td>
<td>2026</td>
<td><code>gpt-5.4-pro</code></td>
<td>Available</td>
<td>1,050,000 tokens</td>
<td>128,000 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2025-08-31</td>
<td><code>medium</code> (default), <code>high</code>, <code>xhigh</code></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.4-pro" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="even">
<td>GPT-5.4 mini</td>
<td>2026</td>
<td><code>gpt-5.4-mini</code></td>
<td>Available</td>
<td>400,000 tokens</td>
<td>128,000 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2025-08-31</td>
<td><code>none</code> (default), <code>low</code>, <code>medium</code>, <code>high</code>, <code>xhigh</code></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.4-mini" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="odd">
<td>GPT-5.4 nano</td>
<td>2026</td>
<td><code>gpt-5.4-nano</code></td>
<td>Available</td>
<td>400,000 tokens</td>
<td>128,000 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2025-08-31</td>
<td><code>none</code> (default), <code>low</code>, <code>medium</code>, <code>high</code>, <code>xhigh</code></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.4-nano" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="even">
<td>GPT-5.5</td>
<td>2026</td>
<td><code>gpt-5.5</code></td>
<td>Available</td>
<td>1,050,000 tokens</td>
<td>128,000 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2025-12-01</td>
<td><code>none</code>, <code>low</code>, <code>medium</code> (default), <code>high</code>, <code>xhigh</code></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.5" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="odd">
<td>GPT-5.5 Pro</td>
<td>2026</td>
<td><code>gpt-5.5-pro</code></td>
<td>Available</td>
<td>1,050,000 tokens</td>
<td>128,000 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2025-12-01</td>
<td><code>medium</code>, <code>high</code> (default), <code>xhigh</code></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.5-pro" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="even">
<td>GPT-5.5 Instant / Chat Latest</td>
<td>2026</td>
<td><code>chat-latest</code></td>
<td>Available<br />
rolling alias</td>
<td>400,000 tokens</td>
<td>128,000 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2025-08-31</td>
<td>—</td>
<td><a href="https://developers.openai.com/api/docs/models/chat-latest" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="odd">
<td>GPT-5.6 Sol</td>
<td>2026</td>
<td><code>gpt-5.6-sol</code><br />
alias: <code>gpt-5.6</code></td>
<td>Available</td>
<td>1,050,000 tokens</td>
<td>128,000 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2026-02-16</td>
<td><code>none</code>, <code>low</code>, <code>medium</code> (default), <code>high</code>, <code>xhigh</code>, <code>max</code>; optional <code>reasoning.mode: "pro"</code></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.6-sol" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="even">
<td>GPT-5.6 Terra</td>
<td>2026</td>
<td><code>gpt-5.6-terra</code></td>
<td>Available</td>
<td>1,050,000 tokens</td>
<td>128,000 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2026-02-16</td>
<td><code>none</code>, <code>low</code>, <code>medium</code> (default), <code>high</code>, <code>xhigh</code>, <code>max</code>; optional <code>reasoning.mode: "pro"</code></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.6-terra" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="odd">
<td>GPT-5.6 Luna</td>
<td>2026</td>
<td><code>gpt-5.6-luna</code></td>
<td>Available</td>
<td>1,050,000 tokens</td>
<td>128,000 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2026-02-16</td>
<td><code>none</code>, <code>low</code>, <code>medium</code> (default), <code>high</code>, <code>xhigh</code>, <code>max</code>; optional <code>reasoning.mode: "pro"</code></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.6-luna" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="even">
<td>GPT-5.6 Cyber</td>
<td>2026</td>
<td><code>gpt-5.6-cyber</code><br />
alias: <code>gpt-daybreak-red-latest</code></td>
<td>Restricted access<br />
separate Daybreak Red approval and provisioning required</td>
<td>400,000 tokens</td>
<td>128,000 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2026-02-16</td>
<td>Reasoning supported; effort levels not publicly specified</td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.6-cyber" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
<tr class="odd">
<td>GPT-6 Astra</td>
<td>2026</td>
<td><code>gpt-6-astra</code></td>
<td>Limited rollout</td>
<td>1,050,000 tokens</td>
<td>128,000 tokens</td>
<td>Text, image</td>
<td>Text</td>
<td>2026-04-30</td>
<td><code>low</code>, <code>medium</code>, <code>high</code>, <code>xhigh</code>, <code>max</code></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-6-astra" class="external text" rel="nofollow">OpenAI API docs</a></td>
</tr>
</tbody>
</table>

GPT models without disclosed architectural parameters (as of 2026-07-10)

## Models

<table class="wikitable sortable" style="width:100%;">
<colgroup>
<col style="width: 14%" />
<col style="width: 14%" />
<col style="width: 14%" />
<col style="width: 14%" />
<col style="width: 14%" />
<col style="width: 14%" />
<col style="width: 14%" />
</colgroup>
<thead>
<tr class="header">
<th>No.</th>
<th>Date</th>
<th>Model</th>
<th>Note</th>
<th>Source</th>
<th>API docs</th>
<th>System / model card</th>
</tr>
</thead>
<tbody>
<tr class="odd">
<td>1</td>
<td>2018-06-11</td>
<td>GPT / GPT-1</td>
<td>First GPT-line model; transformer-based language model trained with unsupervised pre-training followed by supervised fine-tuning.</td>
<td><a href="https://openai.com/index/language-unsupervised/" class="external text" rel="nofollow">OpenAI</a></td>
<td>—</td>
<td>—</td>
</tr>
<tr class="even">
<td>2</td>
<td>2019-02-14</td>
<td>GPT-2</td>
<td>Large unsupervised language model released gradually because of misuse concerns.</td>
<td><a href="https://openai.com/index/better-language-models/" class="external text" rel="nofollow">OpenAI</a></td>
<td>—</td>
<td>—</td>
</tr>
<tr class="odd">
<td>3</td>
<td>2020-05-28</td>
<td>GPT-3</td>
<td>175-billion-parameter language model demonstrating strong few-shot learning.</td>
<td><a href="https://openai.com/index/language-models-are-few-shot-learners/" class="external text" rel="nofollow">OpenAI</a></td>
<td>—</td>
<td>—</td>
</tr>
<tr class="even">
<td>4</td>
<td>2022-01-27</td>
<td>InstructGPT</td>
<td>Instruction-following GPT-3-based models trained with reinforcement learning from human feedback.</td>
<td><a href="https://openai.com/index/instruction-following/" class="external text" rel="nofollow">OpenAI</a></td>
<td>—</td>
<td>—</td>
</tr>
<tr class="odd">
<td>5</td>
<td>2022-11-30</td>
<td>ChatGPT / GPT-3.5</td>
<td>Conversational model related to InstructGPT; launched as a research preview.</td>
<td><a href="https://openai.com/index/chatgpt/" class="external text" rel="nofollow">OpenAI</a></td>
<td>—</td>
<td>—</td>
</tr>
<tr class="even">
<td>6</td>
<td>2023-03-01</td>
<td>GPT-3.5 Turbo</td>
<td>API model family used in ChatGPT; optimized for dialogue and lower-cost inference.</td>
<td><a href="https://openai.com/index/introducing-chatgpt-and-whisper-apis/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-3.5-turbo" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td>—</td>
</tr>
<tr class="odd">
<td>7</td>
<td>2023-03-14</td>
<td>GPT-4</td>
<td>Large multimodal model accepting image and text inputs and producing text outputs.</td>
<td><a href="https://openai.com/index/gpt-4-research/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-4" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td><a href="https://cdn.openai.com/papers/gpt-4-system-card.pdf" class="external text" rel="nofollow">OpenAI PDF</a></td>
</tr>
<tr class="even">
<td>8</td>
<td>2023-11-06</td>
<td>GPT-4 Turbo</td>
<td>More capable and lower-cost GPT-4 variant introduced with a 128K-token context window.</td>
<td><a href="https://openai.com/index/new-models-and-developer-products-announced-at-devday/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-4-turbo" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td>—</td>
</tr>
<tr class="odd">
<td>9</td>
<td>2024-05-13</td>
<td>GPT-4o</td>
<td>Omni model designed for real-time multimodal interaction across text, vision, and audio.</td>
<td><a href="https://openai.com/index/hello-gpt-4o/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-4o" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td><a href="https://openai.com/index/gpt-4o-system-card/" class="external text" rel="nofollow">OpenAI System Card</a></td>
</tr>
<tr class="even">
<td>10</td>
<td>2024-07-18</td>
<td>GPT-4o mini</td>
<td>Smaller and lower-cost GPT-4o-class model for cost-efficient multimodal intelligence.</td>
<td><a href="https://openai.com/index/gpt-4o-mini-advancing-cost-efficient-intelligence/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-4o-mini" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td>Evaluated in the GPT-4o System Card:<br />
<a href="https://openai.com/index/gpt-4o-system-card/" class="external text" rel="nofollow">OpenAI</a></td>
</tr>
<tr class="odd">
<td>11</td>
<td>2025-02-27</td>
<td>GPT-4.5</td>
<td>Research-preview GPT model focused on scaling unsupervised pre-training and improving non-reasoning intelligence.</td>
<td><a href="https://openai.com/index/introducing-gpt-4-5/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-4.5-preview" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td><a href="https://openai.com/index/gpt-4-5-system-card/" class="external text" rel="nofollow">OpenAI System Card</a></td>
</tr>
<tr class="even">
<td>12</td>
<td>2025-04-14</td>
<td>GPT-4.1</td>
<td>API model with improvements in coding, instruction following, and long-context use.</td>
<td><a href="https://openai.com/index/gpt-4-1/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-4.1" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td>—</td>
</tr>
<tr class="odd">
<td>13</td>
<td>2025-04-14</td>
<td>GPT-4.1 mini</td>
<td>Smaller and faster version of GPT-4.1.</td>
<td><a href="https://openai.com/index/gpt-4-1/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-4.1-mini" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td>—</td>
</tr>
<tr class="even">
<td>14</td>
<td>2025-04-14</td>
<td>GPT-4.1 nano</td>
<td>Smallest and lowest-cost version of GPT-4.1.</td>
<td><a href="https://openai.com/index/gpt-4-1/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-4.1-nano" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td>—</td>
</tr>
<tr class="odd">
<td>15</td>
<td>2025-08-05</td>
<td>gpt-oss-120b</td>
<td>Large open-weight reasoning model released under the Apache 2.0 license.</td>
<td><a href="https://openai.com/index/introducing-gpt-oss/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-oss-120b" class="external text" rel="nofollow">OpenAI model docs</a></td>
<td><a href="https://openai.com/index/gpt-oss-model-card/" class="external text" rel="nofollow">OpenAI Model Card</a></td>
</tr>
<tr class="even">
<td>16</td>
<td>2025-08-05</td>
<td>gpt-oss-20b</td>
<td>Smaller open-weight reasoning model released under the Apache 2.0 license.</td>
<td><a href="https://openai.com/index/introducing-gpt-oss/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-oss-20b" class="external text" rel="nofollow">OpenAI model docs</a></td>
<td><a href="https://openai.com/index/gpt-oss-model-card/" class="external text" rel="nofollow">OpenAI Model Card</a></td>
</tr>
<tr class="odd">
<td>17</td>
<td>2025-08-07</td>
<td>GPT-5</td>
<td>Unified flagship model system combining fast responses with adaptive reasoning.</td>
<td><a href="https://openai.com/index/introducing-gpt-5/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td><a href="https://openai.com/index/gpt-5-system-card/" class="external text" rel="nofollow">OpenAI System Card</a></td>
</tr>
<tr class="even">
<td>18</td>
<td>2025-08-07</td>
<td>GPT-5 mini</td>
<td>Smaller GPT-5 API model balancing capability, cost, and latency.</td>
<td><a href="https://openai.com/index/introducing-gpt-5-for-developers/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5-mini" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td>Covered in the GPT-5 System Card:<br />
<a href="https://openai.com/index/gpt-5-system-card/" class="external text" rel="nofollow">OpenAI</a></td>
</tr>
<tr class="odd">
<td>19</td>
<td>2025-08-07</td>
<td>GPT-5 nano</td>
<td>Smallest GPT-5 API model for high-volume, low-latency, and low-cost workloads.</td>
<td><a href="https://openai.com/index/introducing-gpt-5-for-developers/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5-nano" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td>Covered in the GPT-5 System Card:<br />
<a href="https://openai.com/index/gpt-5-system-card/" class="external text" rel="nofollow">OpenAI</a></td>
</tr>
<tr class="even">
<td>20</td>
<td>2025-08-07</td>
<td>GPT-5 Pro / GPT-5 Thinking Pro</td>
<td>Extended-reasoning GPT-5 variant intended for harder and higher-stakes tasks.</td>
<td><a href="https://openai.com/index/introducing-gpt-5/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5-pro" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td>Covered in the GPT-5 System Card:<br />
<a href="https://openai.com/index/gpt-5-system-card/" class="external text" rel="nofollow">OpenAI</a></td>
</tr>
<tr class="odd">
<td>21</td>
<td>2025-09-15</td>
<td>GPT-5-Codex</td>
<td>GPT-5 variant optimized for agentic software-engineering tasks in Codex.</td>
<td><a href="https://openai.com/index/introducing-upgrades-to-codex/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5-codex" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td><a href="https://openai.com/index/gpt-5-system-card-addendum-gpt-5-codex/" class="external text" rel="nofollow">OpenAI System Card Addendum</a></td>
</tr>
<tr class="even">
<td>22</td>
<td>2025-11-12</td>
<td>GPT-5.1 Instant</td>
<td>More conversational GPT-5.1 chat model with improved instruction following.</td>
<td><a href="https://openai.com/index/gpt-5-1/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.1-chat-latest" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td><a href="https://openai.com/index/gpt-5-system-card-addendum-gpt-5-1/" class="external text" rel="nofollow">OpenAI System Card Addendum</a></td>
</tr>
<tr class="odd">
<td>23</td>
<td>2025-11-12</td>
<td>GPT-5.1 Thinking</td>
<td>GPT-5.1 reasoning model with more adaptive allocation of thinking time.</td>
<td><a href="https://openai.com/index/gpt-5-1/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.1" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td><a href="https://openai.com/index/gpt-5-system-card-addendum-gpt-5-1/" class="external text" rel="nofollow">OpenAI System Card Addendum</a></td>
</tr>
<tr class="even">
<td>24</td>
<td>2025-11-13</td>
<td>GPT-5.1-Codex</td>
<td>GPT-5.1 variant optimized for long-running agentic coding and software-engineering tasks.</td>
<td><a href="https://openai.com/index/gpt-5-1-for-developers/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.1-codex" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td>—</td>
</tr>
<tr class="odd">
<td>25</td>
<td>2025-11-13</td>
<td>GPT-5.1-Codex mini</td>
<td>Smaller and lower-cost GPT-5.1-Codex model for agentic coding workloads.</td>
<td><a href="https://openai.com/index/gpt-5-1-for-developers/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.1-codex-mini" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td>—</td>
</tr>
<tr class="even">
<td>26</td>
<td>2025-11-19</td>
<td>GPT-5.1-Codex-Max</td>
<td>Frontier agentic coding model with context compaction for long-running software tasks.</td>
<td><a href="https://openai.com/index/gpt-5-1-codex-max/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.1-codex-max" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td><a href="https://openai.com/index/gpt-5-1-codex-max-system-card/" class="external text" rel="nofollow">OpenAI System Card</a></td>
</tr>
<tr class="odd">
<td>27</td>
<td>2025-12-11</td>
<td>GPT-5.2 Instant</td>
<td>Instant GPT-5.2 variant for everyday ChatGPT interactions and API workloads.</td>
<td><a href="https://openai.com/index/introducing-gpt-5-2/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.2-chat-latest" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td><a href="https://openai.com/index/gpt-5-system-card-update-gpt-5-2/" class="external text" rel="nofollow">OpenAI System Card Update</a></td>
</tr>
<tr class="even">
<td>28</td>
<td>2025-12-11</td>
<td>GPT-5.2 Thinking</td>
<td>Reasoning variant of GPT-5.2 with stronger performance on scientific, mathematical, and professional tasks.</td>
<td><a href="https://openai.com/index/introducing-gpt-5-2/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.2" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td><a href="https://openai.com/index/gpt-5-system-card-update-gpt-5-2/" class="external text" rel="nofollow">OpenAI System Card Update</a></td>
</tr>
<tr class="odd">
<td>29</td>
<td>2025-12-11</td>
<td>GPT-5.2 Pro</td>
<td>Pro variant of GPT-5.2 for harder reasoning and professional work.</td>
<td><a href="https://openai.com/index/introducing-gpt-5-2/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.2-pro" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td><a href="https://openai.com/index/gpt-5-system-card-update-gpt-5-2/" class="external text" rel="nofollow">OpenAI System Card Update</a></td>
</tr>
<tr class="even">
<td>30</td>
<td>2025-12-18</td>
<td>GPT-5.2-Codex</td>
<td>GPT-5.2 variant optimized for long-horizon agentic software engineering and cybersecurity work.</td>
<td><a href="https://openai.com/index/introducing-gpt-5-2-codex/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.2-codex" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td><a href="https://openai.com/index/gpt-5-2-codex-system-card/" class="external text" rel="nofollow">OpenAI System Card</a></td>
</tr>
<tr class="odd">
<td>31</td>
<td>2026-02-05</td>
<td>GPT-5.3-Codex</td>
<td>Specialized agentic coding model for long-running and complex software-engineering tasks.</td>
<td><a href="https://openai.com/index/introducing-gpt-5-3-codex/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.3-codex" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td><a href="https://openai.com/index/gpt-5-3-codex-system-card/" class="external text" rel="nofollow">OpenAI System Card</a></td>
</tr>
<tr class="even">
<td>32</td>
<td>2026-02-12</td>
<td>GPT-5.3-Codex-Spark</td>
<td>Research-preview, text-only coding model optimized for low-latency, real-time interaction in Codex.</td>
<td><a href="https://openai.com/index/introducing-gpt-5-3-codex-spark/" class="external text" rel="nofollow">OpenAI</a></td>
<td>—</td>
<td>—</td>
</tr>
<tr class="odd">
<td>33</td>
<td>2026-03-03</td>
<td>GPT-5.3 Instant</td>
<td>Updated everyday ChatGPT model focused on smoother, more useful, and better-grounded responses.</td>
<td><a href="https://openai.com/index/gpt-5-3-instant/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.3-chat-latest" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td><a href="https://openai.com/index/gpt-5-3-instant-system-card/" class="external text" rel="nofollow">OpenAI System Card</a></td>
</tr>
<tr class="even">
<td>34</td>
<td>2026-03-05</td>
<td>GPT-5.4</td>
<td>General-purpose frontier model with native computer-use capabilities.</td>
<td><a href="https://openai.com/index/introducing-gpt-5-4/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.4" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td><a href="https://openai.com/index/gpt-5-4-thinking-system-card/" class="external text" rel="nofollow">OpenAI System Card</a></td>
</tr>
<tr class="odd">
<td>35</td>
<td>2026-03-05</td>
<td>GPT-5.4 Pro</td>
<td>Pro version of GPT-5.4 for more complex professional, reasoning, and agentic tasks.</td>
<td><a href="https://openai.com/index/introducing-gpt-5-4/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.4-pro" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td>Covered in the GPT-5.4 Thinking System Card:<br />
<a href="https://openai.com/index/gpt-5-4-thinking-system-card/" class="external text" rel="nofollow">OpenAI</a></td>
</tr>
<tr class="even">
<td>36</td>
<td>2026-03-17</td>
<td>GPT-5.4 mini</td>
<td>Smaller and faster GPT-5.4-class model for coding, tool use, and multimodal reasoning.</td>
<td><a href="https://openai.com/index/introducing-gpt-5-4-mini-and-nano/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.4-mini" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td>Covered in an appendix to the GPT-5.4 Thinking System Card:<br />
<a href="https://openai.com/index/gpt-5-4-thinking-system-card/" class="external text" rel="nofollow">OpenAI</a></td>
</tr>
<tr class="odd">
<td>37</td>
<td>2026-03-17</td>
<td>GPT-5.4 nano</td>
<td>Smallest GPT-5.4-class model for high-volume and low-cost API workloads.</td>
<td><a href="https://openai.com/index/introducing-gpt-5-4-mini-and-nano/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.4-nano" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td>—</td>
</tr>
<tr class="even">
<td>38</td>
<td>2026-04-23</td>
<td>GPT-5.5</td>
<td>Flagship GPT-5.5 model for professional work, coding, research, and analysis.</td>
<td><a href="https://openai.com/index/introducing-gpt-5-5/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.5" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td><a href="https://openai.com/index/gpt-5-5-system-card/" class="external text" rel="nofollow">OpenAI System Card</a></td>
</tr>
<tr class="odd">
<td>39</td>
<td>2026-04-23</td>
<td>GPT-5.5 Pro</td>
<td>Pro version of GPT-5.5 for more demanding reasoning and professional tasks.</td>
<td><a href="https://openai.com/index/introducing-gpt-5-5/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.5-pro" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td>Covered in the GPT-5.5 System Card:<br />
<a href="https://openai.com/index/gpt-5-5-system-card/" class="external text" rel="nofollow">OpenAI</a></td>
</tr>
<tr class="even">
<td>40</td>
<td>2026-05-05</td>
<td>GPT-5.5 Instant</td>
<td>Updated default ChatGPT model with improved accuracy, clarity, and personalization.</td>
<td><a href="https://openai.com/index/gpt-5-5-instant/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/chat-latest" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td><a href="https://openai.com/index/gpt-5-5-instant-system-card/" class="external text" rel="nofollow">OpenAI System Card</a></td>
</tr>
<tr class="odd">
<td>41</td>
<td>2026-06-26</td>
<td>GPT-5.6 Sol</td>
<td>Flagship GPT-5.6 tier for complex professional work. Entered limited preview on 2026-06-26 and became generally available on 2026-07-09.</td>
<td><a href="https://openai.com/index/previewing-gpt-5-6-sol/" class="external text" rel="nofollow">OpenAI preview announcement</a><br />
<a href="https://openai.com/index/gpt-5-6/" class="external text" rel="nofollow">OpenAI general-release announcement</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.6-sol" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td><a href="https://deploymentsafety.openai.com/gpt-5-6" class="external text" rel="nofollow">OpenAI Deployment Safety Hub</a></td>
</tr>
<tr class="even">
<td>42</td>
<td>2026-06-26</td>
<td>GPT-5.6 Terra</td>
<td>Balanced GPT-5.6 tier for everyday professional work. Entered limited preview on 2026-06-26 and became generally available on 2026-07-09.</td>
<td><a href="https://openai.com/index/previewing-gpt-5-6-sol/" class="external text" rel="nofollow">OpenAI preview announcement</a><br />
<a href="https://openai.com/index/gpt-5-6/" class="external text" rel="nofollow">OpenAI general-release announcement</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.6-terra" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td><a href="https://deploymentsafety.openai.com/gpt-5-6" class="external text" rel="nofollow">OpenAI Deployment Safety Hub</a></td>
</tr>
<tr class="odd">
<td>43</td>
<td>2026-06-26</td>
<td>GPT-5.6 Luna</td>
<td>Fastest and lowest-cost GPT-5.6 tier. Entered limited preview on 2026-06-26 and became generally available on 2026-07-09.</td>
<td><a href="https://openai.com/index/previewing-gpt-5-6-sol/" class="external text" rel="nofollow">OpenAI preview announcement</a><br />
<a href="https://openai.com/index/gpt-5-6/" class="external text" rel="nofollow">OpenAI general-release announcement</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.6-luna" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td><a href="https://deploymentsafety.openai.com/gpt-5-6" class="external text" rel="nofollow">OpenAI Deployment Safety Hub</a></td>
</tr>
<tr class="even">
<td>44</td>
<td>2026-07-09</td>
<td>GPT-5.6 Sol Pro</td>
<td>Higher-compute product mode based on GPT-5.6 Sol for the most difficult reasoning and professional tasks; not a separately documented base API model.</td>
<td><a href="https://openai.com/index/gpt-5-6/" class="external text" rel="nofollow">OpenAI</a></td>
<td>—</td>
<td>Covered in the GPT-5.6 deployment-safety materials:<br />
<a href="https://deploymentsafety.openai.com/gpt-5-6" class="external text" rel="nofollow">OpenAI Deployment Safety Hub</a></td>
</tr>
<tr class="odd">
<td>45</td>
<td>2026-08-10</td>
<td>GPT-5.6 Cyber</td>
<td>Specialized GPT-5.6 model for approved professionals conducting advanced authorized cybersecurity research, vulnerability discovery, exploit validation, and defensive security testing.</td>
<td><a href="https://openai.com/index/expanding-daybreak-as-the-cyber-defense-window-narrows/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-5.6-cyber" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td>Covered in the GPT-5.6 deployment-safety materials:<br />
<a href="https://deploymentsafety.openai.com/gpt-5-6" class="external text" rel="nofollow">OpenAI Deployment Safety Hub</a></td>
</tr>
<tr class="even">
<td>46</td>
<td>2026-09-03</td>
<td>GPT-6 Astra</td>
<td>First publicly announced GPT-6-generation model and new flagship for the hardest end-to-end work, with advances in computer use, browsing, software engineering, cybersecurity, scientific research, and long-horizon agentic workflows.</td>
<td><a href="https://openai.com/index/gpt-6-astra/" class="external text" rel="nofollow">OpenAI</a></td>
<td><a href="https://developers.openai.com/api/docs/models/gpt-6-astra" class="external text" rel="nofollow">OpenAI API docs</a></td>
<td><a href="https://openai.com/index/safety-overview-gpt-6-astra/" class="external text" rel="nofollow">OpenAI Safety Overview</a></td>
</tr>
<tr class="odd">
<td>47</td>
<td>2026-09-03</td>
<td>GPT-6 Pro</td>
<td>Higher-compute ChatGPT product mode powered by GPT-6 Astra for the most difficult and longer-running tasks; introduced as part of the staged GPT-6 rollout.</td>
<td><a href="https://openai.com/index/gpt-6-astra/" class="external text" rel="nofollow">OpenAI</a></td>
<td>—</td>
<td>Covered in the GPT-6 Astra safety overview:<br />
<a href="https://openai.com/index/safety-overview-gpt-6-astra/" class="external text" rel="nofollow">OpenAI</a></td>
</tr>
</tbody>
</table>

## Benchmarks

As of 10 July 2026, this table records selected benchmark results reported directly by OpenAI. It is a source-indexed record of published evaluations, not a controlled leaderboard. Only evaluation runs for which OpenAI reported at least one result in the selected columns are included. Models with no reported result in these columns are omitted rather than shown as wholly blank rows.

The date in the first column is the publication date of the evaluation source, not necessarily the model's release date. A model may appear more than once when OpenAI evaluated a different snapshot, changed the benchmark protocol, reasoning setting, tool access, task set, or harness, or published a materially different later result. Scores are percentages unless otherwise stated. An em dash (—) means that the cited source did not report that benchmark for that exact evaluation run; it does not mean that no result exists elsewhere.

<table class="wikitable sortable" style="text-align:center; width:100%;">
<caption>Selected OpenAI-reported benchmark results</caption>
<colgroup>
<col style="width: 8%" />
<col style="width: 8%" />
<col style="width: 8%" />
<col style="width: 8%" />
<col style="width: 8%" />
<col style="width: 8%" />
<col style="width: 8%" />
<col style="width: 8%" />
<col style="width: 8%" />
<col style="width: 8%" />
<col style="width: 8%" />
<col style="width: 8%" />
</colgroup>
<thead>
<tr class="header">
<th>Evaluation published</th>
<th>Evaluated model / snapshot</th>
<th>Variant / evaluation setup<sup>a</sup></th>
<th><a href="https://systems-analysis.info/eng/MMLU_Benchmark" title="MMLU Benchmark">MMLU</a><sup>b</sup></th>
<th><a href="https://systems-analysis.info/eng/GSM8K_(Grade_School_Math_8K)" title="GSM8K (Grade School Math 8K)">GSM8K</a><sup>c</sup></th>
<th><a href="https://systems-analysis.info/eng/HumanEval_Benchmark" title="HumanEval Benchmark">HumanEval</a><sup>d</sup></th>
<th><a href="https://systems-analysis.info/eng/MATH_Benchmark" title="MATH Benchmark">MATH</a><sup>e</sup></th>
<th><a href="https://systems-analysis.info/eng/GPQA_Diamond_Benchmark" title="GPQA Diamond Benchmark">GPQA-D</a><sup>f</sup></th>
<th><a href="https://systems-analysis.info/eng/MMMU_Benchmark" title="MMMU Benchmark">MMMU</a><sup>g</sup></th>
<th><a href="https://systems-analysis.info/eng/MMMU-Pro_Benchmark" title="MMMU-Pro Benchmark">MMMU-Pro</a><sup>h</sup></th>
<th><a href="https://systems-analysis.info/eng/SWE-bench_Verified" title="SWE-bench Verified">SWE-V</a><sup>i</sup></th>
<th><a href="https://systems-analysis.info/eng/SWE-bench_Pro" title="SWE-bench Pro">SWE-Pro</a><sup>j</sup></th>
</tr>
</thead>
<tbody>
<tr class="odd">
<td>2023-03-27</td>
<td>GPT-3.5<br />
GPT-4 Technical Report comparator<sup>k</sup></td>
<td>MMLU and GSM8K 5-shot; HumanEval 0-shot</td>
<td>70.0</td>
<td>57.1</td>
<td>48.1</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
</tr>
<tr class="even">
<td>2023-03-27</td>
<td>GPT-4 base model<br />
GPT-4 Technical Report evaluation<sup>k</sup></td>
<td>MMLU 5-shot; GSM8K 5-shot chain-of-thought; HumanEval 0-shot</td>
<td>86.4</td>
<td>92.0</td>
<td>67.0</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
</tr>
<tr class="odd">
<td>2024-05-13</td>
<td>GPT-4<br />
GPT-4o launch comparison<sup>l</sup></td>
<td>Text-evaluation comparison; source label: “GPT-4 (Initial release 23-03-14)”</td>
<td>86.4</td>
<td>—</td>
<td>67.0</td>
<td>42.5</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
</tr>
<tr class="even">
<td>2024-05-13</td>
<td>GPT-4 Turbo<sup>l</sup></td>
<td>GPT-4o launch text-evaluation comparison</td>
<td>86.5</td>
<td>—</td>
<td>87.1</td>
<td>72.6</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
</tr>
<tr class="odd">
<td>2024-05-13</td>
<td>GPT-4o<br />
launch evaluation<sup>l</sup></td>
<td>GPT-4o launch text-evaluation comparison</td>
<td>88.7</td>
<td>—</td>
<td>90.2</td>
<td>76.6</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
</tr>
<tr class="even">
<td>2024-07-18</td>
<td>GPT-4o mini<br />
launch evaluation<sup>m</sup></td>
<td>OpenAI <code>simple-evals</code>; API assistant system-message prompt</td>
<td>82.0</td>
<td>—</td>
<td>87.2</td>
<td>—</td>
<td>—</td>
<td>59.4</td>
<td>—</td>
<td>—</td>
<td>—</td>
</tr>
<tr class="odd">
<td>2025-04-14</td>
<td>GPT-4o (<code>2024-11-20</code>)<sup>n</sup></td>
<td>GPT-4.1 appendix; GPQA extraction caveat; SWE-V <code>n=477</code></td>
<td>85.7</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>46.0</td>
<td>68.7</td>
<td>—</td>
<td>33.2</td>
<td>—</td>
</tr>
<tr class="even">
<td>2025-04-14</td>
<td>GPT-4o mini<br />
GPT-4.1 comparison<sup>n</sup></td>
<td>GPT-4.1 appendix; GPQA extraction caveat; SWE-V <code>n=477</code></td>
<td>82.0</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>40.2</td>
<td>56.3</td>
<td>—</td>
<td>8.7</td>
<td>—</td>
</tr>
<tr class="odd">
<td>2025-04-14</td>
<td>GPT-4.5<br />
GPT-4.1 comparison<sup>n</sup></td>
<td>GPT-4.1 appendix; GPQA extraction caveat; SWE-V <code>n=477</code></td>
<td>90.8</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>69.5</td>
<td>75.2</td>
<td>—</td>
<td>38.0</td>
<td>—</td>
</tr>
<tr class="even">
<td>2025-04-14</td>
<td>GPT-4.1<sup>n</sup></td>
<td>GPT-4.1 appendix; GPQA extraction caveat; SWE-V <code>n=477</code></td>
<td>90.2</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>66.3</td>
<td>74.8</td>
<td>—</td>
<td>54.6</td>
<td>—</td>
</tr>
<tr class="odd">
<td>2025-04-14</td>
<td>GPT-4.1 mini<sup>n</sup></td>
<td>GPT-4.1 appendix; GPQA extraction caveat; SWE-V <code>n=477</code></td>
<td>87.5</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>65.0</td>
<td>72.7</td>
<td>—</td>
<td>23.6</td>
<td>—</td>
</tr>
<tr class="even">
<td>2025-04-14</td>
<td>GPT-4.1 nano<sup>n</sup></td>
<td>GPT-4.1 appendix; GPQA extraction caveat</td>
<td>80.1</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>50.3</td>
<td>55.4</td>
<td>—</td>
<td>—</td>
<td>—</td>
</tr>
<tr class="odd">
<td>2025-08-05</td>
<td>gpt-oss-120b<sup>o</sup></td>
<td>High reasoning; default system prompt; GPQA-D without tools; SWE-V <code>n=477</code></td>
<td>90.0</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>80.1</td>
<td>—</td>
<td>—</td>
<td>62.4</td>
<td>—</td>
</tr>
<tr class="even">
<td>2025-08-05</td>
<td>gpt-oss-20b<sup>o</sup></td>
<td>High reasoning; default system prompt; GPQA-D without tools; SWE-V <code>n=477</code></td>
<td>85.3</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>71.5</td>
<td>—</td>
<td>—</td>
<td>60.7</td>
<td>—</td>
</tr>
<tr class="odd">
<td>2025-08-07</td>
<td>GPT-5<br />
API launch evaluation<sup>p</sup></td>
<td>High reasoning; GPQA-D without tools; SWE-V <code>n=477</code></td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>85.7</td>
<td>84.2</td>
<td>78.4</td>
<td>74.9</td>
<td>—</td>
</tr>
<tr class="even">
<td>2025-08-07</td>
<td>GPT-5 mini<br />
API launch evaluation<sup>p</sup></td>
<td>High reasoning; GPQA-D without tools; SWE-V <code>n=477</code></td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>82.3</td>
<td>81.6</td>
<td>74.1</td>
<td>71.0</td>
<td>—</td>
</tr>
<tr class="odd">
<td>2025-08-07</td>
<td>GPT-5 nano<br />
API launch evaluation<sup>p</sup></td>
<td>High reasoning; GPQA-D without tools; SWE-V <code>n=477</code></td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>71.2</td>
<td>75.6</td>
<td>62.6</td>
<td>54.7</td>
<td>—</td>
</tr>
<tr class="even">
<td>2025-08-07</td>
<td>GPT-5 Pro<sup>p</sup></td>
<td>Extended reasoning; GPQA-D without tools</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>88.4</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
</tr>
<tr class="odd">
<td>2025-09-15</td>
<td>GPT-5-Codex<sup>q</sup></td>
<td>High reasoning; all 500 SWE-V tasks</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>74.5</td>
<td>—</td>
</tr>
<tr class="even">
<td>2025-11-13</td>
<td>GPT-5.1<br />
API evaluation<sup>r</sup></td>
<td>High reasoning; GPQA-D without tools; all 500 SWE-V tasks; JSON-based <code>apply_patch</code> harness</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>88.1</td>
<td>85.4</td>
<td>—</td>
<td>76.3</td>
<td>—</td>
</tr>
<tr class="odd">
<td>2025-11-13</td>
<td>GPT-5<br />
GPT-5.1 comparison re-evaluation<sup>r</sup></td>
<td>High reasoning; GPQA-D without tools; all 500 SWE-V tasks; JSON-based <code>apply_patch</code> harness</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>85.7</td>
<td>84.2</td>
<td>—</td>
<td>72.8</td>
<td>—</td>
</tr>
<tr class="even">
<td>2025-11-19</td>
<td>GPT-5.1-Codex<sup>s</sup></td>
<td>High reasoning; all 500 SWE-V tasks</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>73.7</td>
<td>—</td>
</tr>
<tr class="odd">
<td>2025-11-19</td>
<td>GPT-5.1-Codex-Max<sup>s</sup></td>
<td><code>xhigh</code> reasoning; all 500 SWE-V tasks</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>77.9</td>
<td>—</td>
</tr>
<tr class="even">
<td>2025-12-11</td>
<td>GPT-5.1 Thinking<br />
GPT-5.2 comparison<sup>t</sup></td>
<td>High reasoning; research environment; GPQA-D without tools</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>88.1</td>
<td>—</td>
<td>—</td>
<td>76.3</td>
<td>50.8</td>
</tr>
<tr class="odd">
<td>2025-12-11</td>
<td>GPT-5.2 Thinking<sup>t</sup></td>
<td><code>xhigh</code> reasoning; research environment; MMMU-Pro without tools</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>92.4</td>
<td>—</td>
<td>79.5</td>
<td>80.0</td>
<td>55.6</td>
</tr>
<tr class="even">
<td>2025-12-11</td>
<td>GPT-5.2 Pro<sup>t</sup></td>
<td><code>xhigh</code> reasoning; research environment; GPQA-D without tools</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>93.2</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
</tr>
<tr class="odd">
<td>2026-02-05</td>
<td>GPT-5.2-Codex<br />
GPT-5.3-Codex comparison<sup>u</sup></td>
<td><code>xhigh</code> reasoning; Codex evaluation setup</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>56.4</td>
</tr>
<tr class="even">
<td>2026-02-05</td>
<td>GPT-5.3-Codex<br />
launch evaluation<sup>u</sup></td>
<td><code>xhigh</code> reasoning; Codex evaluation setup</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>56.8</td>
</tr>
<tr class="odd">
<td>2026-03-05</td>
<td>GPT-5.3-Codex<br />
GPT-5.4 comparison re-evaluation<sup>v</sup></td>
<td><code>xhigh</code> reasoning; research environment</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>92.6</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>56.8</td>
</tr>
<tr class="even">
<td>2026-03-05</td>
<td>GPT-5.4<sup>v</sup></td>
<td><code>xhigh</code> reasoning; research environment; MMMU-Pro without tools</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>92.8</td>
<td>—</td>
<td>81.2</td>
<td>—</td>
<td>57.7</td>
</tr>
<tr class="odd">
<td>2026-03-05</td>
<td>GPT-5.4 Pro<sup>v</sup></td>
<td><code>xhigh</code> reasoning; research environment</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>94.4</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
</tr>
<tr class="even">
<td>2026-03-17</td>
<td>GPT-5.4<br />
mini/nano comparison re-evaluation<sup>w</sup></td>
<td><code>xhigh</code> reasoning; plain MMMU-Pro result, not the Python-enabled result</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>93.0</td>
<td>—</td>
<td>81.2</td>
<td>—</td>
<td>57.7</td>
</tr>
<tr class="odd">
<td>2026-03-17</td>
<td>GPT-5.4 mini<sup>w</sup></td>
<td><code>xhigh</code> reasoning; plain MMMU-Pro result, not the Python-enabled result</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>88.0</td>
<td>—</td>
<td>76.6</td>
<td>—</td>
<td>54.4</td>
</tr>
<tr class="even">
<td>2026-03-17</td>
<td>GPT-5.4 nano<sup>w</sup></td>
<td><code>xhigh</code> reasoning; plain MMMU-Pro result, not the Python-enabled result</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>82.8</td>
<td>—</td>
<td>66.1</td>
<td>—</td>
<td>52.4</td>
</tr>
<tr class="odd">
<td>2026-03-17</td>
<td>GPT-5 mini<br />
GPT-5.4 mini/nano comparison re-evaluation<sup>w</sup></td>
<td>High reasoning; plain MMMU-Pro result, not the Python-enabled result</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>81.6</td>
<td>—</td>
<td>67.5</td>
<td>—</td>
<td>45.7</td>
</tr>
<tr class="even">
<td>2026-04-23</td>
<td>GPT-5.5<br />
launch evaluation<sup>x</sup></td>
<td><code>xhigh</code> reasoning; research environment; MMMU-Pro without tools</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>93.6</td>
<td>—</td>
<td>81.2</td>
<td>—</td>
<td>58.6</td>
</tr>
<tr class="odd">
<td>2026-07-09</td>
<td>GPT-5.5<br />
GPT-5.6 comparison re-evaluation<sup>y</sup></td>
<td>Reasoning effort not stated for these entries; MMMU-Pro without tools; source label: “SWE-Bench Pro”</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>93.6</td>
<td>—</td>
<td>81.2</td>
<td>—</td>
<td>59.4</td>
</tr>
<tr class="even">
<td>2026-07-09</td>
<td>GPT-5.6 Sol<sup>y</sup></td>
<td>Reasoning effort not stated for these entries; MMMU-Pro without tools; source label: “SWE-Bench Pro”</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>94.6</td>
<td>—</td>
<td>83.0</td>
<td>—</td>
<td>64.6</td>
</tr>
<tr class="odd">
<td>2026-07-09</td>
<td>GPT-5.6 Terra<sup>y</sup></td>
<td>Reasoning effort not stated for these entries; MMMU-Pro without tools; source label: “SWE-Bench Pro”</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>92.9</td>
<td>—</td>
<td>80.7</td>
<td>—</td>
<td>63.4</td>
</tr>
<tr class="even">
<td>2026-07-09</td>
<td>GPT-5.6 Luna<sup>y</sup></td>
<td>Reasoning effort not stated for these entries; MMMU-Pro without tools; source label: “SWE-Bench Pro”</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>—</td>
<td>92.3</td>
<td>—</td>
<td>78.4</td>
<td>—</td>
<td>62.7</td>
</tr>
</tbody>
</table>

Selected OpenAI-reported benchmark results

**Table notes**

- <sup>a</sup> “Variant / evaluation setup” records the product or API variant, reasoning setting, tool access, snapshot, harness, or evaluated task set stated by the source. Terms such as “Turbo”, “mini”, “nano”, “Instant”, “Thinking”, “Pro”, “Codex”, “high”, and `xhigh` are product or evaluation labels, not disclosed architectural classes. When OpenAI did not identify an exact checkpoint or reasoning setting, the table does not infer one.

<!-- -->

- <sup>b</sup> MMLU entries are limited to results explicitly reported as classic MMLU. Multilingual variants such as MMMLU are not entered in this column. Prompting, answer extraction, checkpoints, and evaluation implementations differ between publications, so the values are not necessarily directly comparable.

<!-- -->

- <sup>c</sup> GSM8K is not interchangeable with MGSM. In particular, the 87.0 result reported for GPT-4o mini at launch is MGSM and is therefore not entered in the GSM8K column.

<!-- -->

- <sup>d</sup> HumanEval values may depend on sampling, pass@k, prompting, contamination controls, and code-execution methodology. Only values explicitly reported as HumanEval are entered.

<!-- -->

- <sup>e</sup> MATH means the original MATH benchmark where the source uses that label. MATH-500, AIME, HMMT, FrontierMath, and other mathematical evaluations are not entered in this column.

<!-- -->

- <sup>f</sup> GPQA-D means GPQA Diamond. Results labelled only “GPQA”, without an explicit indication that the Diamond subset was used, are omitted. Tool access, reasoning effort, answer-extraction methods, and later re-evaluations materially affect comparability.

<!-- -->

- <sup>g</sup> MMMU is the original MMMU benchmark. It is kept separate from MMMU-Pro because the two are different evaluations.

<!-- -->

- <sup>h</sup> MMMU-Pro values are entered separately from classic MMMU. Where a source reports both a no-tools or plain result and a Python- or tool-enabled result, this table uses the no-tools or plain result and identifies the setup in the row or source note.

<!-- -->

- <sup>i</sup> SWE-V means SWE-bench Verified. These figures are preserved as historical OpenAI-reported results, not as a current controlled ranking. Task subsets, denominators, scaffolds, patch tools, retry policies, and the treatment of unexecutable tasks differ between releases. OpenAI later concluded that SWE-bench Verified had substantial test-design and contamination problems and no longer provided a meaningful signal for frontier software-engineering capabilities.<sup>[\[39\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-swev-audit-39)</sup>

<!-- -->

- <sup>j</sup> SWE-Pro covers results that OpenAI labels “SWE-Bench Pro”, “SWE-Bench Pro (Public)”, or “SWE-Bench Pro, Public”; the exact source label is noted where relevant. These figures should also be treated as historical reported results rather than as a clean cross-model ranking. In a July 2026 audit of the 731-task public split, OpenAI's analysis pipeline identified 200 tasks (27.4%) as broken and its human annotation campaign identified 249 tasks (34.1%); OpenAI estimated that approximately 30% of SWE-Bench Pro tasks were broken.<sup>[\[40\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-swepro-audit-40)</sup>

<!-- -->

- <sup>k</sup> The GPT-4 Technical Report evaluates GPT-3.5 as a comparator and a pre-trained base GPT-4 model; the GPT-3.5 row should not be interpreted as a benchmark of the November 2022 ChatGPT product. The report gives MMLU = 70.0 / 86.4, GSM8K = 57.1 / 92.0, and HumanEval = 48.1 / 67.0 for GPT-3.5 / GPT-4. MMLU was evaluated 5-shot and HumanEval 0-shot; GPT-4 GSM8K used 5-shot chain-of-thought prompting. OpenAI mixed training-set data from MATH and GSM8K into GPT-4 pre-training and recommended interpreting the GSM8K result as between true few-shot transfer and benchmark-specific tuning. The report estimated HumanEval contamination at 25%; GPT-4 scored 65.58 on the non-contaminated subset, compared with the headline 67.0.<sup>[\[41\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt4-report-41)</sup>

<!-- -->

- <sup>l</sup> The GPT-4, GPT-4 Turbo, and launch GPT-4o values are transcribed from the “Text Evaluation” comparison published with GPT-4o. The source reports MMLU = 86.4 / 86.5 / 88.7, HumanEval = 67.0 / 87.1 / 90.2, and MATH = 42.5 / 72.6 / 76.6 for GPT-4 / GPT-4 Turbo / GPT-4o. Its science comparison uses the label “GPQA” rather than explicitly identifying GPQA Diamond, so those values are not entered here.<sup>[\[42\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt4o-launch-42)</sup>

<!-- -->

- <sup>m</sup> GPT-4o mini launch values were computed with OpenAI's `simple-evals` repository and the API assistant system-message prompt. The launch page reports MMLU = 82.0, HumanEval = 87.2, and MMMU = 59.4. Its mathematical result of 87.0 is MGSM, not GSM8K.<sup>[\[43\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt4o-mini-43)</sup>

<!-- -->

- <sup>n</sup> GPT-4o (`2024-11-20`), GPT-4o mini, GPT-4.5, and the GPT-4.1 family are taken from the GPT-4.1 comparison appendix. OpenAI's displayed table lists GPT-4o GPQA Diamond = 46.0. The accompanying footnote says that model-based answer extraction can raise GPT-4o from approximately 46 to 54, while changing GPT-4.1 by less than one percentage point; this table reproduces the displayed 46.0 rather than substituting the footnote estimate. For SWE-bench Verified, OpenAI omitted 23 of 500 tasks that could not run on its infrastructure. GPT-4.1 scored 54.6 on the remaining 477 tasks; treating the omitted tasks as zero gives 52.1.<sup>[\[44\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt41-44)</sup>

<!-- -->

- <sup>o</sup> gpt-oss values are taken from the gpt-oss-120b and gpt-oss-20b Model Card. OpenAI reports basic pass@1 results at high reasoning using the default system prompt. GPQA Diamond values are the no-tools results, and SWE-bench Verified uses a 477-task subset.<sup>[\[45\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt-oss-card-45)</sup>

<!-- -->

- <sup>p</sup> GPT-5, GPT-5 mini, and GPT-5 nano launch values use high reasoning. GPQA Diamond was run without tools. The reported MMMU-Pro result is OpenAI's average across its standard and vision sets. SWE-bench Verified omits 23 of 500 tasks and therefore uses 477 tasks. The separate GPT-5 Pro GPQA result used extended reasoning without tools.<sup>[\[46\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt5-dev-46)[\[47\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt5-launch-47)</sup>

<!-- -->

- <sup>q</sup> The GPT-5-Codex launch comparison reports SWE-bench Verified = 74.5 for GPT-5-Codex at high reasoning and 72.8 for GPT-5 at high reasoning, using all 500 tasks. The later GPT-5 comparison result is represented separately because it is not the same 477-task run as the original GPT-5 launch result of 74.9.<sup>[\[48\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt5-codex-48)</sup>

<!-- -->

- <sup>r</sup> GPT-5.1 and its GPT-5 comparison baseline were evaluated at high reasoning. SWE-bench Verified accuracy was averaged over all 500 tasks using a JSON-based `apply_patch` harness, producing 76.3 for GPT-5.1 and 72.8 for GPT-5. The same appendix reports GPQA Diamond = 88.1 / 85.7 and MMMU = 85.4 / 84.2 for GPT-5.1 / GPT-5.

<!-- -->

- <sup>s</sup> GPT-5.1-Codex used high reasoning and scored 73.7 on all 500 SWE-bench Verified tasks. GPT-5.1-Codex-Max used `xhigh` reasoning and scored 77.9 on the same headline evaluation.<sup>[\[49\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt51-codex-max-49)</sup>

<!-- -->

- <sup>t</sup> GPT-5.2 Thinking and GPT-5.2 Pro were evaluated with the maximum reasoning effort available in the API, `xhigh`; the GPT-5.1 Thinking comparison used `high`. OpenAI states that the benchmarks were conducted in a research environment and may differ slightly from production ChatGPT. GPQA Diamond and the displayed MMMU-Pro result are the no-tools results. The source labels the coding benchmark “SWE-Bench Pro, Public”.<sup>[\[50\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt52-50)</sup>

<!-- -->

- <sup>u</sup> The GPT-5.3-Codex launch appendix reports SWE-Bench Pro (Public) = 56.8 for GPT-5.3-Codex and 56.4 for GPT-5.2-Codex. The comparison table labels both models `xhigh`.<sup>[\[51\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt53-codex-51)</sup>

<!-- -->

- <sup>v</sup> GPT-5.4 publication values were produced at `xhigh` reasoning in a research environment. The publication also supplies a later comparison result for GPT-5.3-Codex, including GPQA Diamond = 92.6. GPT-5.4 MMMU-Pro = 81.2 is the no-tools result.<sup>[\[52\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt54-52)</sup>

<!-- -->

- <sup>w</sup> The GPT-5.4 mini and nano comparison evaluates GPT-5.4, GPT-5.4 mini, and GPT-5.4 nano at `xhigh`; GPT-5 mini was evaluated at its maximum available setting, `high`. The table uses the source's plain MMMU-Pro row rather than its separate Python-enabled row. This later comparison reports GPT-5.4 GPQA Diamond = 93.0, compared with 92.8 in the original GPT-5.4 publication, and is therefore retained as a separate evaluation run.<sup>[\[53\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt54-mini-53)</sup>

<!-- -->

- <sup>x</sup> GPT-5.5 launch results were produced at `xhigh` reasoning in a research environment. MMMU-Pro = 81.2 is the no-tools result. The launch table labels the coding benchmark “SWE-Bench Pro (Public)” and reports 58.6. OpenAI also notes evidence of benchmark memorization reported by other laboratories.<sup>[\[54\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt55-54)</sup>

<!-- -->

- <sup>y</sup> The GPT-5.6 general-release evaluation table reports GPQA Diamond = 94.6 / 92.9 / 92.3, MMMU-Pro without tools = 83.0 / 80.7 / 78.4, and SWE-Bench Pro = 64.6 / 63.4 / 62.7 for GPT-5.6 Sol / Terra / Luna. Its GPT-5.5 comparison row gives 93.6 / 81.2 / 59.4 on the same displayed benchmarks. The table does not identify the exact reasoning effort for these entries, so no effort level is inferred. The source uses the label “SWE-Bench Pro”, without “Public”; the later GPT-5.5 result of 59.4 is retained separately from its launch-time result of 58.6.<sup>[\[55\]](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_note-gpt56-55)</sup>

## External links

- “Improving language understanding with unsupervised learning”, <a href="https://openai.com/index/language-unsupervised/" class="external text" rel="nofollow">OpenAI, June 11, 2018</a>
- “Better language models and their implications”, <a href="https://openai.com/index/better-language-models/" class="external text" rel="nofollow">OpenAI, February 14, 2019</a>
- “GPT-2: 6-month follow-up”, <a href="https://openai.com/index/gpt-2-6-month-follow-up/" class="external text" rel="nofollow">OpenAI, August 20, 2019</a>
- “GPT-2: 1.5B release”, <a href="https://openai.com/index/gpt-2-1-5b-release/" class="external text" rel="nofollow">OpenAI, November 5, 2019</a>
- “Language models are few-shot learners”, <a href="https://openai.com/index/language-models-are-few-shot-learners/" class="external text" rel="nofollow">OpenAI, May 28, 2020</a>
- “OpenAI API”, <a href="https://openai.com/index/openai-api/" class="external text" rel="nofollow">OpenAI, June 11, 2020</a>
- “Introducing ChatGPT”, <a href="https://openai.com/index/chatgpt/" class="external text" rel="nofollow">OpenAI, November 30, 2022</a>
- “Function calling and other API updates”, <a href="https://openai.com/index/function-calling-and-other-api-updates/" class="external text" rel="nofollow">OpenAI, June 13, 2023</a>
- “GPT-4”, <a href="https://openai.com/index/gpt-4-research/" class="external text" rel="nofollow">OpenAI, March 14, 2023</a>
- “New models and developer products announced at DevDay”, <a href="https://openai.com/index/new-models-and-developer-products-announced-at-devday/" class="external text" rel="nofollow">OpenAI, November 6, 2023</a>
- “Hello GPT-4o”, <a href="https://openai.com/index/hello-gpt-4o/" class="external text" rel="nofollow">OpenAI, May 13, 2024</a>
- “Introducing GPT-4.1 in the API”, <a href="https://openai.com/index/gpt-4-1/" class="external text" rel="nofollow">OpenAI, April 14, 2025</a>
- “Introducing GPT-4.5”, <a href="https://openai.com/index/introducing-gpt-4-5/" class="external text" rel="nofollow">OpenAI, February 27, 2025</a>
- “Introducing GPT-5”, <a href="https://openai.com/index/introducing-gpt-5/" class="external text" rel="nofollow">OpenAI, August 7, 2025</a>
- “GPT-5.1: A smarter, more conversational ChatGPT”, <a href="https://openai.com/index/gpt-5-1/" class="external text" rel="nofollow">OpenAI, November 12, 2025</a>
- “Introducing GPT-5.2”, <a href="https://openai.com/index/introducing-gpt-5-2/" class="external text" rel="nofollow">OpenAI, December 11, 2025</a>
- “Introducing GPT-5.2-Codex”, <a href="https://openai.com/index/introducing-gpt-5-2-codex/" class="external text" rel="nofollow">OpenAI, December 18, 2025</a>
- “Introducing GPT-5.3-Codex”, <a href="https://openai.com/index/introducing-gpt-5-3-codex/" class="external text" rel="nofollow">OpenAI, February 5, 2026</a>
- “Introducing GPT-5.3-Codex-Spark”, <a href="https://openai.com/index/introducing-gpt-5-3-codex-spark/" class="external text" rel="nofollow">OpenAI, February 12, 2026</a>
- “GPT-5.3 Instant: Smoother, more useful everyday conversations”, <a href="https://openai.com/index/gpt-5-3-instant/" class="external text" rel="nofollow">OpenAI, March 3, 2026</a>
- “Introducing GPT-5.4”, <a href="https://openai.com/index/introducing-gpt-5-4/" class="external text" rel="nofollow">OpenAI, March 5, 2026</a>
- “Using GPT-5.4”, <a href="https://developers.openai.com/api/docs/guides/latest-model/" class="external text" rel="nofollow">OpenAI Developers</a>
- “Models”, <a href="https://developers.openai.com/api/docs/models" class="external text" rel="nofollow">OpenAI API</a>
- “Deprecations”, <a href="https://developers.openai.com/api/docs/deprecations/" class="external text" rel="nofollow">OpenAI API</a>
- “GPT-5.3 and GPT-5.4 in ChatGPT”, <a href="https://help.openai.com/en/articles/11909943-gpt-53-and-54-in-chatgpt" class="external text" rel="nofollow">OpenAI Help Center</a>
- “Retiring GPT-4o and other ChatGPT models”, <a href="https://help.openai.com/en/articles/20001051-retiring-gpt-4o-and-other-chatgpt-models" class="external text" rel="nofollow">OpenAI Help Center</a>
- <a href="https://en.wikipedia.org/wiki/Generative_pre-trained_transformer" class="external text" rel="nofollow">Generative pre-trained transformer — Wikipedia</a>

## Bibliography

- Radford, A. et al. (2018). *Improving Language Understanding by Generative Pre-Training*. <a href="https://cdn.openai.com/research-covers/language-unsupervised/language_understanding_paper.pdf" class="external text" rel="nofollow">PDF</a>.
- Radford, A. et al. (2019). *Language Models are Unsupervised Multitask Learners*. <a href="https://cdn.openai.com/better-language-models/language_models_are_unsupervised_multitask_learners.pdf" class="external text" rel="nofollow">PDF</a>.
- Brown, T. B. et al. (2020). *Language Models are Few-Shot Learners*. <a href="https://arxiv.org/abs/2005.14165" class="external text" rel="nofollow">arXiv:2005.14165</a>.
- Kaplan, J. et al. (2020). *Scaling Laws for Neural Language Models*. <a href="https://arxiv.org/abs/2001.08361" class="external text" rel="nofollow">arXiv:2001.08361</a>.
- Chen, M. et al. (2021). *Evaluating Large Language Models Trained on Code*. <a href="https://arxiv.org/abs/2107.03374" class="external text" rel="nofollow">arXiv:2107.03374</a>.
- Ouyang, L. et al. (2022). *Training Language Models to Follow Instructions with Human Feedback*. <a href="https://arxiv.org/abs/2203.02155" class="external text" rel="nofollow">arXiv:2203.02155</a>.
- Hoffmann, J. et al. (2022). *Training Compute-Optimal Large Language Models*. <a href="https://arxiv.org/abs/2203.15556" class="external text" rel="nofollow">arXiv:2203.15556</a>.
- Bai, Y. et al. (2022). *Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback*. <a href="https://arxiv.org/abs/2204.05862" class="external text" rel="nofollow">arXiv:2204.05862</a>.
- OpenAI (2023). *GPT-4 Technical Report*. <a href="https://arxiv.org/abs/2303.08774" class="external text" rel="nofollow">arXiv:2303.08774</a>.
- Bubeck, S. et al. (2023). *Sparks of Artificial General Intelligence: Early Experiments with GPT-4*. <a href="https://arxiv.org/abs/2303.12712" class="external text" rel="nofollow">arXiv:2303.12712</a>.

## References

1.  <span id="cite_note-gpt5-announce-1">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt5-announce_1-0) OpenAI. "Introducing GPT-5" (August 7, 2025). <a href="https://openai.com/index/introducing-gpt-5/" class="external free" rel="nofollow">https://openai.com/index/introducing-gpt-5/</a></span>
2.  <span id="cite_note-2">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-2) OpenAI. "Introducing GPT-4.5" (2025). <a href="https://openai.com/index/introducing-gpt-4-5/" class="external free" rel="nofollow">https://openai.com/index/introducing-gpt-4-5/</a></span>
3.  <span id="cite_note-gpt45-systemcard-3">↑ <sup>[3.0](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt45-systemcard_3-0)</sup> <sup>[3.1](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt45-systemcard_3-1)</sup> OpenAI. GPT-4.5 System Card (February 27, 2025). <a href="https://cdn.openai.com/gpt-4-5-system-card-2272025.pdf" class="external free" rel="nofollow">https://cdn.openai.com/gpt-4-5-system-card-2272025.pdf</a></span>
4.  <span id="cite_note-deprecations-4">↑ <sup>[4.0](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-deprecations_4-0)</sup> <sup>[4.1](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-deprecations_4-1)</sup> OpenAI Developers. Deprecations. <a href="https://developers.openai.com/api/docs/deprecations/" class="external free" rel="nofollow">https://developers.openai.com/api/docs/deprecations/</a></span>
5.  <span id="cite_note-5">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-5) OpenAI. "Introducing GPT-4.1 in the API" (2025).</span>
6.  <span id="cite_note-gpt53-instant-6">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt53-instant_6-0) OpenAI. "GPT-5.3 Instant: Smoother, more useful everyday conversations" (March 3, 2026). <a href="https://openai.com/index/gpt-5-3-instant/" class="external free" rel="nofollow">https://openai.com/index/gpt-5-3-instant/</a></span>
7.  <span id="cite_note-gpt54-announce-7">↑ <sup>[7.0](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt54-announce_7-0)</sup> <sup>[7.1](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt54-announce_7-1)</sup> <sup>[7.2](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt54-announce_7-2)</sup> OpenAI. "Introducing GPT-5.4" (March 5, 2026). <a href="https://openai.com/index/introducing-gpt-5-4/" class="external free" rel="nofollow">https://openai.com/index/introducing-gpt-5-4/</a></span>
8.  <span id="cite_note-gpt54-api-8">↑ <sup>[8.0](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt54-api_8-0)</sup> <sup>[8.1](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt54-api_8-1)</sup> OpenAI Developers. "Using GPT-5.4". <a href="https://developers.openai.com/api/docs/guides/latest-model/" class="external free" rel="nofollow">https://developers.openai.com/api/docs/guides/latest-model/</a></span>
9.  <span id="cite_note-radford2018-9">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-radford2018_9-0) Radford, A. et al. (2018). *Improving Language Understanding by Generative Pre-Training*. <a href="https://cdn.openai.com/research-covers/language-unsupervised/language_understanding_paper.pdf" class="external free" rel="nofollow">https://cdn.openai.com/research-covers/language-unsupervised/language_understanding_paper.pdf</a></span>
10. <span id="cite_note-10">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-10) OpenAI noted that the updated GPT-3.5 Turbo "now comes by default with 16k context."</span>
11. <span id="cite_note-11">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-11) OpenAI. "GPT-4 Technical Report" (2023). arXiv:2303.08774.</span>
12. <span id="cite_note-12">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-12) These estimates are based on data published by SemiAnalysis and corroborated by a number of independent sources.</span>
13. <span id="cite_note-13">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-13) Announced at OpenAI DevDay on November 6, 2023; general availability from April 9, 2024.</span>
14. <span id="cite_note-14">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-14) The codename Orion and the characterization "last model without chain-of-thought" appeared in Sam Altman's roadmap communications and several media publications (Reuters, The Verge), but not in the GPT-4.5 launch post itself.</span>
15. <span id="cite_note-15">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-15) OpenAI. "Introducing GPT-4.1 in the API" (2025).</span>
16. <span id="cite_note-16">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-16) OpenAI. "Introducing GPT-5" (August 7, 2025).</span>
17. <span id="cite_note-chatgpt-53-54-17">↑ <sup>[17.0](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-chatgpt-53-54_17-0)</sup> <sup>[17.1](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-chatgpt-53-54_17-1)</sup> <sup>[17.2](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-chatgpt-53-54_17-2)</sup> OpenAI Help Center. "GPT-5.3 and GPT-5.4 in ChatGPT". <a href="https://help.openai.com/en/articles/11909943-gpt-53-and-54-in-chatgpt" class="external free" rel="nofollow">https://help.openai.com/en/articles/11909943-gpt-53-and-54-in-chatgpt</a></span>
18. <span id="cite_note-18">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-18) OpenAI API documentation. Models: GPT-5.</span>
19. <span id="cite_note-19">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-19) OpenAI. "Introducing GPT-5" (2025). Testing conditions: "with web search enabled on anonymized prompts representative of ChatGPT production traffic."</span>
20. <span id="cite_note-gpt51-api-20">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt51-api_20-0) OpenAI Developers. Models: GPT-5.1. <a href="https://developers.openai.com/api/docs/models/gpt-5.1" class="external free" rel="nofollow">https://developers.openai.com/api/docs/models/gpt-5.1</a></span>
21. <span id="cite_note-gpt51-announce-21">↑ <sup>[21.0](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt51-announce_21-0)</sup> <sup>[21.1](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt51-announce_21-1)</sup> <sup>[21.2](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt51-announce_21-2)</sup> OpenAI. "GPT-5.1: A smarter, more conversational ChatGPT" (November 12, 2025). <a href="https://openai.com/index/gpt-5-1/" class="external free" rel="nofollow">https://openai.com/index/gpt-5-1/</a></span>
22. <span id="cite_note-gpt51-dev-22">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt51-dev_22-0) OpenAI. "GPT-5.1 for developers" (2025). <a href="https://openai.com/index/gpt-5-1-for-developers/" class="external free" rel="nofollow">https://openai.com/index/gpt-5-1-for-developers/</a></span>
23. <span id="cite_note-23">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-23) OpenAI. "Introducing GPT-5.3-Codex" (February 5, 2026). <a href="https://openai.com/index/introducing-gpt-5-3-codex/" class="external free" rel="nofollow">https://openai.com/index/introducing-gpt-5-3-codex/</a></span>
24. <span id="cite_note-24">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-24) OpenAI. "Introducing GPT-5.3-Codex-Spark" (February 12, 2026). <a href="https://openai.com/index/introducing-gpt-5-3-codex-spark/" class="external free" rel="nofollow">https://openai.com/index/introducing-gpt-5-3-codex-spark/</a></span>
25. <span id="cite_note-gpt54-model-25">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt54-model_25-0) OpenAI Developers. Models: GPT-5.4. <a href="https://developers.openai.com/api/docs/models/gpt-5.4" class="external free" rel="nofollow">https://developers.openai.com/api/docs/models/gpt-5.4</a></span>
26. <span id="cite_note-gpt55-announce-26">↑ <sup>[26.0](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt55-announce_26-0)</sup> <sup>[26.1](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt55-announce_26-1)</sup> OpenAI. *Introducing GPT-5.5*. <a href="https://openai.com/index/introducing-gpt-5-5/" class="external free" rel="nofollow">https://openai.com/index/introducing-gpt-5-5/</a></span>
27. <span id="cite_note-gpt55-api-27">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt55-api_27-0) OpenAI Developers. Models: GPT-5.5. <a href="https://developers.openai.com/api/docs/models/gpt-5.5" class="external free" rel="nofollow">https://developers.openai.com/api/docs/models/gpt-5.5</a></span>
28. <span id="cite_note-gpt55-pro-api-28">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt55-pro-api_28-0) OpenAI Developers. Models: GPT-5.5 Pro. <a href="https://developers.openai.com/api/docs/models/gpt-5.5-pro" class="external free" rel="nofollow">https://developers.openai.com/api/docs/models/gpt-5.5-pro</a></span>
29. <span id="cite_note-gpt55-instant-29">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt55-instant_29-0) OpenAI. *GPT-5.5 Instant: smarter, clearer, and more personalized*. <a href="https://openai.com/index/gpt-5-5-instant/" class="external free" rel="nofollow">https://openai.com/index/gpt-5-5-instant/</a></span>
30. <span id="cite_note-gpt56-announce-30">↑ <sup>[30.0](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt56-announce_30-0)</sup> <sup>[30.1](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt56-announce_30-1)</sup> <sup>[30.2](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt56-announce_30-2)</sup> OpenAI. *GPT-5.6: Frontier intelligence that scales with your ambition*. <a href="https://openai.com/index/gpt-5-6/" class="external free" rel="nofollow">https://openai.com/index/gpt-5-6/</a></span>
31. <span id="cite_note-gpt56-sol-api-31">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt56-sol-api_31-0) OpenAI Developers. Models: GPT-5.6 Sol. <a href="https://developers.openai.com/api/docs/models/gpt-5.6-sol" class="external free" rel="nofollow">https://developers.openai.com/api/docs/models/gpt-5.6-sol</a></span>
32. <span id="cite_note-gpt56-terra-api-32">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt56-terra-api_32-0) OpenAI Developers. Models: GPT-5.6 Terra. <a href="https://developers.openai.com/api/docs/models/gpt-5.6-terra" class="external free" rel="nofollow">https://developers.openai.com/api/docs/models/gpt-5.6-terra</a></span>
33. <span id="cite_note-gpt56-luna-api-33">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt56-luna-api_33-0) OpenAI Developers. Models: GPT-5.6 Luna. <a href="https://developers.openai.com/api/docs/models/gpt-5.6-luna" class="external free" rel="nofollow">https://developers.openai.com/api/docs/models/gpt-5.6-luna</a></span>
34. <span id="cite_note-gpt6-announce-34">↑ <sup>[34.0](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt6-announce_34-0)</sup> <sup>[34.1](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt6-announce_34-1)</sup> OpenAI. *GPT-6 Astra: A new generation of intelligence*. September 3, 2026.</span>
35. <span id="cite_note-gpt6-guide-35">↑ <sup>[35.0](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt6-guide_35-0)</sup> <sup>[35.1](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt6-guide_35-1)</sup> OpenAI Developers. *Model guidance: Using GPT-6 Astra*.</span>
36. <span id="cite_note-gpt6-api-36">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt6-api_36-0) OpenAI Developers. *Models: GPT-6 Astra*.</span>
37. <span id="cite_note-gpt6-chatgpt-37">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt6-chatgpt_37-0) OpenAI Help Center. *GPT-5.6 and GPT-6 Pro in ChatGPT*.</span>
38. <span id="cite_note-gpt6-safety-38">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt6-safety_38-0) OpenAI. *Safety overview: GPT-6 Astra*. September 3, 2026.</span>
39. <span id="cite_note-swev-audit-39">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-swev-audit_39-0) OpenAI. *Why SWE-bench Verified no longer measures frontier coding capabilities*. 23 February 2026. <a href="https://openai.com/index/why-we-no-longer-evaluate-swe-bench-verified/" class="external text" rel="nofollow">OpenAI</a>. Accessed 10 July 2026.</span>
40. <span id="cite_note-swepro-audit-40">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-swepro-audit_40-0) OpenAI. *Separating signal from noise in coding evaluations*. 8 July 2026. <a href="https://openai.com/index/separating-signal-from-noise-coding-evaluations/" class="external text" rel="nofollow">OpenAI</a>. Accessed 10 July 2026.</span>
41. <span id="cite_note-gpt4-report-41">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt4-report_41-0) OpenAI. *GPT-4 Technical Report*. 27 March 2023. <a href="https://cdn.openai.com/papers/gpt-4.pdf" class="external text" rel="nofollow">PDF</a>. Accessed 10 July 2026.</span>
42. <span id="cite_note-gpt4o-launch-42">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt4o-launch_42-0) OpenAI. *Hello GPT-4o*. 13 May 2024. <a href="https://openai.com/index/hello-gpt-4o/" class="external text" rel="nofollow">OpenAI</a>. Accessed 10 July 2026.</span>
43. <span id="cite_note-gpt4o-mini-43">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt4o-mini_43-0) OpenAI. *GPT-4o mini: advancing cost-efficient intelligence*. 18 July 2024. <a href="https://openai.com/index/gpt-4o-mini-advancing-cost-efficient-intelligence/" class="external text" rel="nofollow">OpenAI</a>. Accessed 10 July 2026.</span>
44. <span id="cite_note-gpt41-44">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt41_44-0) OpenAI. *Introducing GPT-4.1 in the API*. 14 April 2025. <a href="https://openai.com/index/gpt-4-1/" class="external text" rel="nofollow">OpenAI</a>. Accessed 10 July 2026.</span>
45. <span id="cite_note-gpt-oss-card-45">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt-oss-card_45-0) OpenAI. *gpt-oss-120b & gpt-oss-20b Model Card*. 5 August 2025. <a href="https://cdn.openai.com/pdf/419b6906-9da6-406c-a19d-1bb078ac7637/oai_gpt-oss_model_card.pdf" class="external text" rel="nofollow">PDF</a>. Accessed 10 July 2026.</span>
46. <span id="cite_note-gpt5-dev-46">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt5-dev_46-0) OpenAI. *Introducing GPT-5 for developers*. 7 August 2025. <a href="https://openai.com/index/introducing-gpt-5-for-developers/" class="external text" rel="nofollow">OpenAI</a>. Accessed 10 July 2026.</span>
47. <span id="cite_note-gpt5-launch-47">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt5-launch_47-0) OpenAI. *Introducing GPT-5*. 7 August 2025. <a href="https://openai.com/index/introducing-gpt-5/" class="external text" rel="nofollow">OpenAI</a>. Accessed 10 July 2026.</span>
48. <span id="cite_note-gpt5-codex-48">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt5-codex_48-0) OpenAI. *Introducing upgrades to Codex*. 15 September 2025. <a href="https://openai.com/index/introducing-upgrades-to-codex/" class="external text" rel="nofollow">OpenAI</a>. Accessed 10 July 2026.</span>
49. <span id="cite_note-gpt51-codex-max-49">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt51-codex-max_49-0) OpenAI. *Building more with GPT-5.1-Codex-Max*. 19 November 2025. <a href="https://openai.com/index/gpt-5-1-codex-max/" class="external text" rel="nofollow">OpenAI</a>. Accessed 10 July 2026.</span>
50. <span id="cite_note-gpt52-50">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt52_50-0) OpenAI. *Introducing GPT-5.2*. 11 December 2025. <a href="https://openai.com/index/introducing-gpt-5-2/" class="external text" rel="nofollow">OpenAI</a>. Accessed 10 July 2026.</span>
51. <span id="cite_note-gpt53-codex-51">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt53-codex_51-0) OpenAI. *Introducing GPT-5.3-Codex*. 5 February 2026. <a href="https://openai.com/index/introducing-gpt-5-3-codex/" class="external text" rel="nofollow">OpenAI</a>. Accessed 10 July 2026.</span>
52. <span id="cite_note-gpt54-52">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt54_52-0) OpenAI. *Introducing GPT-5.4*. 5 March 2026. <a href="https://openai.com/index/introducing-gpt-5-4/" class="external text" rel="nofollow">OpenAI</a>. Accessed 10 July 2026.</span>
53. <span id="cite_note-gpt54-mini-53">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt54-mini_53-0) OpenAI. *Introducing GPT-5.4 mini and nano*. 17 March 2026. <a href="https://openai.com/index/introducing-gpt-5-4-mini-and-nano/" class="external text" rel="nofollow">OpenAI</a>. Accessed 10 July 2026.</span>
54. <span id="cite_note-gpt55-54">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt55_54-0) OpenAI. *Introducing GPT-5.5*. 23 April 2026. <a href="https://openai.com/index/introducing-gpt-5-5/" class="external text" rel="nofollow">OpenAI</a>. Accessed 10 July 2026.</span>
55. <span id="cite_note-gpt56-55">[↑](https://systems-analysis.info/eng/GPT_(OpenAI)#cite_ref-gpt56_55-0) OpenAI. *GPT-5.6: Frontier intelligence that scales with your ambition*. 9 July 2026. <a href="https://openai.com/index/gpt-5-6/" class="external text" rel="nofollow">OpenAI</a>. Accessed 10 July 2026.</span>