| Benchmark | Description | Gemini 3.7 Flash | Gemini 3.6 Flash | Claude Sonnet 5 | GPT-5.6 Terra | Muse Spark 1.2 | |
|---|---|---|---|---|---|---|---|
| Input price | $/M tokens | Lower is better | $0.75* | $0.75* | $2.00 | $2.00 | $1.25 |
| Output price | $/M tokens | Lower is better | $3.75* | $3.75* | $10.00 | $12.00 | $4.25 |
| Artificial Analysis Intelligence Index |
Composite model intelligence | 56 | 52 | 55 | 57 | 57 | |
| FrontierCode 11 | Main production code quality | 46.6% | 34.4% | 42.7% | 41.3% | — | |
| DeepSWE v1.1 | Long-form software engineering | 65.3% | 48.6% | 53.8% | 69.6% | 54.9% | |
| Code Arena | Web development | 1,588 | 1,538 | 1,541 | 1,523 | 1,535 | |
| Terminal-Bench 2.1 | Agentic terminal coding | 85.8% | 78.0% | 80.4% | 87.4% | 82.9% | |
| Terminal-Bench 3.0 | General agent capabilities | 14.9% | 5.4% | 14.6% | 20.8% | — | |
| AutomationBench | Enterprise workflow automation | 30.4% | 17.0% | 10.7% | 23.6% | — | |
| GDPval-AA v2 | Knowledge tools | 1,525 | 1,422 | 1,598 | 1,578 | 1,628 | |
| Harvey LAB-AA | Complex logic workflows | 90.7% | 85.1% | 90.1% | 85.2% | — | |
| DocBench (GOP) | Expert PDF document comprehension | 34.0% | 22.0% | 28.0% | 24.7% | 16.0% | |
| CharXiv Reasoning | Information synthesis from complex charts |
No tools With tools |
84.5% 88.7% |
85.2% 89.4% |
77.0% 88.3% |
85.9% — |
— — |
| LVBench | Long video understanding | 85.4% | 84.2% | 68.5% | 78.9% | — | |
| MRCR v2 (B-needle) | Long context performance | 97.0% | 91.8% | 81.5% | 93.5% | — | |
| OSWorld-2.0 | Agentic computer use | 47.9% | 33.8% | — | 50.2% | — | |
| Agent's Last Exam | Multimodal desktop and OS agent tests |
26.3% | 24.2% | 33.3% | 28.0% | — | |
| HLE-Verified | Multidisciplinary expert reasoning | 53.6% | 51.2% | 31.0% | 51.1% | — | |
| BioMysteryBench | Bioinformatics research reasoning | Overall Human difficult |
87.1% 43.5% |
80.6% 41.2% |
87.5% 34.1% |
83.8% 49.4% |
— — |
| LABBench2 | Biology web world research tests | 82.1% | 76.1% | 80.1% | 81.2% | — |