--- title: "YandexGPT (language model) (ID)" source: "https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)" wiki: "systems-analysis.info/int" article: "YandexGPT_(language_model)_(ID)" language: "id" categories: - "Category:Indonesian" - "Category:Large language models" - "Category:LLM families" - "Category:Machine learning" revision_id: 8519 wiki_created_at: 2026-09-07T01:18:34Z wiki_modified_at: 2026-09-07T01:18:34Z downloaded_at: 2026-09-07T23:25:49Z --- # YandexGPT (language model) (ID) **YandexGPT (Yet another GPT)** — keluarga model bahasa besar yang dikembangkan oleh Yandex dan pertama kali diperkenalkan pada Mei 2023.^{[\[1\]](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_note-rbc-may17-1)} Jaringan YandexGPT digunakan dalam asisten bawaan Alice, Pencarian, dan layanan lainnya, serta tersedia melalui API publik platform Yandex Cloud.^{[\[2\]](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_note-cloud-api-2)} **YaLM‑100B** (2022) — model riset pendahulu dengan 100 miliar parameter dan kode terbuka; model ini berfungsi sebagai "bukti konsep", namun YandexGPT dikembangkan secara terpisah untuk penggunaan komersial.^{[\[3\]](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_note-github-yalm-3)} ## Riwayat Rilis | Tanggal | Rilis | Fitur Utama | |-------------|---------------------------|----------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | Jun 2022 | YaLM‑100B | 100 B parameter, 1,7 TB data; Apache 2.0.^{[\[3\]](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_note-github-yalm-3)} | | 17 Mei 2023 | YandexGPT 1.0 | Integrasi ke dalam "Alice".^{[\[1\]](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_note-rbc-may17-1)} | | 7 Sep 2023 | YandexGPT 2 | +67% kualitas berdasarkan uji internal.^{[\[4\]](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_note-rbc-sep8-4)} | | 28 Mar 2024 | YandexGPT 3 Pro / Lite | Lini API korporat baru.^{[\[5\]](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_note-rbc-mar28-5)} | | 24 Okt 2024 | YandexGPT 4 Pro / Lite | Konteks 32.000 token; penalaran tersembunyi (chain‑of‑thought).^{[\[6\]](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_note-habr-gpt4-6)} | | 25 Feb 2025 | YandexGPT 5 Pro | Paritas dengan GPT‑4o pada 64% tugas.^{[\[7\]](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_note-adindex-5pro-7)} | | 31 Mar 2025 | YandexGPT 5 Lite Instruct | Model 8 miliar parameter dengan akses terbuka; format Llama.^{[\[8\]](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_note-hf-5lite-8)} | Versi-versi Utama ## Arsitektur dan Pelatihan - Arsitektur dasar: transformer yang dioptimalkan untuk bahasa Rusia. - YandexGPT 5 Lite: kompatibel dengan Llama; pre-training ≈ 15 triliun token, fine-tuning lanjutan ≈ 320 miliar.^{[\[8\]](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_note-hf-5lite-8)} ### Konteks dan Batas - Batas konteks arsitektural — 32.000 token (versi 4/5).^{[\[6\]](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_note-habr-gpt4-6)} - API publik membatasi satu permintaan (prompt + completion) sebesar 7.400 token.^{[\[9\]](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_note-langchain-7400-9)} - Ukuran maksimum \*\*respons\*\* — 2.000 token sesuai bagian "Quotas and limits".^{[\[10\]](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_note-cloud-limits-10)} ## Model Saat Ini (Juni 2025) | Model | Parameter | Konteks | Lisensi | Keterangan | |------------------|------------|---------|-------------------------|--------------------------------------------------------------------------------------------------------------------------------------------------| | YandexGPT 5 Pro | n/a | 32.000 | proprietari | Akses melalui API dan "Alice Pro".^{[\[7\]](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_note-adindex-5pro-7)} | | YandexGPT 5 Lite | 8 miliar | 32.000 | Yandex GPT‑Lite License | Terbuka; kompatibel dengan Llama.^{[\[8\]](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_note-hf-5lite-8)} | | YaLM‑100B | 100 miliar | 2.048 | Apache 2.0 | Proyek awal.^{[\[3\]](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_note-github-yalm-3)} | ## Benchmark - Uji internal: 5 Pro mencapai paritas dengan GPT‑4o pada 64% tugas; peningkatan dibanding 4 Pro — 67%.^{[\[7\]](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_note-adindex-5pro-7)} - ru‑LLM Arena: YandexGPT mempertahankan kepemimpinan berdasarkan peringkat ELO di antara model berbahasa Rusia.^{[\[11\]](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_note-llmarena-11)} ## Fine‑tuning Untuk 5 Lite, metode LoRA didukung secara resmi; contoh penggunaan dipublikasikan dalam model card.^{[\[8\]](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_note-hf-5lite-8)} ## Mode API - Sinkron — respons cepat (Lite). - Asinkron — tugas-tugas yang intensif sumber daya (Pro).^{[\[2\]](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_note-cloud-api-2)} ## Multimodalitas Keluarga YandexGPT tetap berbasis teks; layanan multimodal ("Neuro", "YandexArt", "Yandex Vision") dikembangkan secara terpisah.^{[\[6\]](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_note-habr-gpt4-6)} ## Tautan - YandexGPT di Yandex Cloud — halaman layanan - Bobot YandexGPT‑5 Lite di Hugging Face - Repositori YaLM‑100B di GitHub ## Literatur - Matkin, N. et al. (2024). *Comparative Analysis of Encoder-Based NER and Large Language Models for Skill Extraction from Russian Job Vacancies*. arXiv:2407.19816. - Tsanda, A.; Bruches, E. (2024). *Russian-Language Multimodal Dataset for Automatic Summarization of Scientific Papers*. arXiv:2405.07886. - Goloburda, M. et al. (2025). *Qorǵau: Evaluating LLM Safety in Kazakh-Russian Bilingual Contexts*. arXiv:2502.13640. - Togmanov, M. et al. (2025). *KazMMLU: Evaluating Language Models on Kazakh, Russian, and Regional Knowledge of Kazakhstan*. arXiv:2502.12829. - Noels, S. et al. (2025). *What Large Language Models Do Not Talk About: An Empirical Study of Moderation and Censorship Practices*. arXiv:2504.03803. ## Catatan 1. ↑ ^{[1.0](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_ref-rbc-may17_1-0)} ^{[1.1](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_ref-rbc-may17_1-1)} «“Яндекс” добавил в “Алису” аналог ChatGPT». *РБК*. [1] 2. ↑ ^{[2.0](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_ref-cloud-api_2-0)} ^{[2.1](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_ref-cloud-api_2-1)} «Getting started with YandexGPT (Quickstart)». *Yandex Cloud Docs*. [2] 3. ↑ ^{[3.0](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_ref-github-yalm_3-0)} ^{[3.1](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_ref-github-yalm_3-1)} ^{[3.2](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_ref-github-yalm_3-2)} «yandex/YaLM‑100B: Pretrained language model with 100B». *GitHub*. [3] 4. [↑](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_ref-rbc-sep8_4-0) «Как “Яндекс” решил зарабатывать на своём аналоге ChatGPT». *РБК*. [4] 5. [↑](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_ref-rbc-mar28_5-0) «“Яндекс” представил третье поколение нейросетей YandexGPT». *РБК*. [5] 6. ↑ ^{[6.0](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_ref-habr-gpt4_6-0)} ^{[6.1](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_ref-habr-gpt4_6-1)} ^{[6.2](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_ref-habr-gpt4_6-2)} «Более мощное семейство моделей YandexGPT 4». *Habr*. [6] 7. ↑ ^{[7.0](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_ref-adindex-5pro_7-0)} ^{[7.1](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_ref-adindex-5pro_7-1)} ^{[7.2](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_ref-adindex-5pro_7-2)} «“Яндекс” внедрил YandexGPT 5 Pro в чат с “Алисой Про”». *AdIndex*. [7] 8. ↑ ^{[8.0](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_ref-hf-5lite_8-0)} ^{[8.1](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_ref-hf-5lite_8-1)} ^{[8.2](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_ref-hf-5lite_8-2)} ^{[8.3](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_ref-hf-5lite_8-3)} «yandex/YandexGPT‑5‑Lite‑8B‑pretrain». *Hugging Face*. [8] 9. [↑](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_ref-langchain-7400_9-0) «ChatYandexGPT API Reference (max_tokens = 7400)». *LangChain Docs*. [9] 10. [↑](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_ref-cloud-limits_10-0) «Yandex Cloud service quotas and limits → Foundation Models». *Yandex Cloud Docs*. [10] 11. [↑](https://systems-analysis.info/int/YandexGPT_(language_model)_(ID)#cite_ref-llmarena_11-0) «llmarena/llmarena — российская краудсорсинговая платформа оценки LLM». *GitHub*. [11]