Google
availableShows if the model has enough results for an index.Gemini 3.7 Flash
Gemini 3.7 Flash is a reasoning model from Google. 51 benchmarks count toward its score, in 8 categories.
IndexOverall score out of 100.72.3 ±2.3
CoverageShare of the index weight with results.100%
SpeedOutput tokens per second.80/s
Input / 1MUS dollars per 1M input tokens.$0.75 batch $0.375
Output / 1MUS dollars per 1M output tokens.$3.75 batch $1.88 US dollars per 1M output tokens in a batch.
ContextMaximum tokens in one request.1.05M
EloLMArena rating and rank.1490 (#8)
The index is a score out of 100. The ± range shows how much it can change. Batch work costs less.
5,640 votes. Elo shows what people prefer. It does not change the score.
CapabilitiesScore per category, out of 100.
Out of 100Results
51 counted| BenchmarkThe test name. | CategoryThe capability that the test measures. | ResultThe score from the publisher. | IndexThis result as a score out of 100. | RunThe settings of the run. | DateDate of the result. | Published byThe source of the result. |
|---|---|---|---|---|---|---|
| OTIS Mock AIME 2024-2025 | Math | 97.2% | 65.9 | high effort | — | Epoch AI |
| OpenAI MRCR v2 8-needle 64K-128K | Reasoning | 97.0% | — | — | — | OpenAI |
| ARC-AGI-1 (semi-private) | Reasoning | 95.5% | 72.3 | high effort | — | ARC Prize Foundation |
| GPQA diamond | Knowledge | 94.8% | 65.8 | high effort | — | Epoch AI |
| Artificial Analysis GPQA Diamond | Knowledge | 94.5% | 65.4 | — | — | Artificial Analysis |
| GPQA Diamond | Knowledge | 93.9% | 65.0 | — | 1 Sept 2026 | Vals AI |
| LiveBench Mathematics | Math | 93.5% | 72.8 | high effort | 25 Jun 2026 | LiveBench |
| Artificial Analysis Harvey LAB-AA | Agentic | 90.7% | 73.6 | — | — | Artificial Analysis |
| MMLU Pro | Knowledge | 90.1% | 62.5 | — | 1 Sept 2026 | Vals AI |
| MMMU Pro | Multimodal | 89.0% | 72.1 | — | 1 Sept 2026 | Vals AI |
| CharXiv Reasoning | Multimodal | 88.7% | 67.1 | — | — | CharXiv authors |
| LiveCodeBench | Coding | 88.7% | 64.9 | — | 1 Sept 2026 | Vals AI |
| LiveBench Reasoning | Reasoning | 87.8% | 77.9 | high effort | 25 Jun 2026 | LiveBench |
| BioMysteryBench Human Solvable | Knowledge | 87.1% | — | — | — | Anthropic |
| Terminal-Bench 2.1 (provider run) | Agentic | 85.8% | 74.7 | — | — | DeepSeek-AI |
| Terminal-Bench 2.1 (provider run) | Agentic | 85.8% | 74.7 | — | — | DeepSeek-AI |
| Artificial Analysis MMMU-Pro | Multimodal | 85.5% | 71.6 | — | — | Artificial Analysis |
| LiveBench Language | Knowledge | 85.5% | 73.6 | high effort | 25 Jun 2026 | LiveBench |
| LVBench | Multimodal | 85.4% | — | — | — | Qwen Team |
| ARC-AGI-2 (semi-private) | Reasoning | 84.6% | 82.4 | high effort | — | ARC Prize Foundation |
| CharXiv Reasoning without tools | Multimodal | 84.5% | — | — | — | CharXiv authors |
| LABBench2: An Improved Benchmark for AI Systems Performing Biology Research | Knowledge | 82.1% | — | — | — | Jon M. Laurent et al. |
| Artificial Analysis Long Context Reasoning | Reasoning | 81.7% | 64.8 | — | — | Artificial Analysis |
| SWE-bench | Coding | 80.8% | 62.4 | — | 1 Sept 2026 | Vals AI |
| LiveBench Instruction Following | Instruction | 79.9% | 86.0 | high effort | 25 Jun 2026 | LiveBench |
| LiveBench Coding | Coding | 78.9% | 68.6 | high effort | 25 Jun 2026 | LiveBench |
| Terminal-Bench 2.1 | Agentic | 77.5% | 69.8 | — | 21 Sept 2026 | Vals AI |
| EuroEval French | Multilingual | 77.1% | 95.0 | — | — | EuroEval |
| EuroEval Italian | Multilingual | 76.9% | 95.0 | — | — | EuroEval |
| EuroEval Swedish | Multilingual | 76.6% | 95.0 | — | — | EuroEval |
| Artificial Analysis Coding Index | Coding | 76.1% | 72.6 | — | — | Artificial Analysis |
| EuroEval Portuguese | Multilingual | 74.4% | 95.0 | — | — | EuroEval |
| EuroEval Dutch | Multilingual | 73.1% | 93.5 | — | — | EuroEval |
| FrontierMath-Tiers-1-3-v2-Private | Math | 71.6% | 71.7 | high effort | — | Epoch AI |
| Vibe Code Bench v1.1 | Coding | 70.4% | 71.5 | OpenHands | 21 Sept 2026 | Vals AI |
| SimpleQA Verified | Knowledge | 69.2% | 85.6 | high effort | — | Epoch AI |
| EuroEval Spanish | Multilingual | 68.3% | 87.5 | — | — | EuroEval |
| LiveBench Data Analysis | Reasoning | 68.0% | 50.4 | high effort | 25 Jun 2026 | LiveBench |
| EuroEval Polish | Multilingual | 67.9% | 87.0 | — | — | EuroEval |
| IOI | Coding | 67.8% | 75.2 | — | 21 Sept 2026 | Vals AI |
| EuroEval German | Multilingual | 67.3% | 86.3 | — | — | EuroEval |
| SkillsBench | Coding | 65.9% | 78.8 | OpenHands | 11 Sept 2026 | Vals AI |
| DeepSWE | Agentic | 65.3% | 71.2 | — | — | Datacurve AI |
| Artificial Analysis AnalystAgent | Agentic | 60.0% | 83.6 | — | — | Artificial Analysis |
| LiveBench Agentic Coding | Agentic | 58.3% | 71.6 | high effort | 25 Jun 2026 | LiveBench |
| ProofBench v1.1 | Math | 58.0% | 72.9 | — | 21 Sept 2026 | Vals AI |
| Artificial Analysis SciCode | Coding | 57.2% | 72.2 | — | — | Artificial Analysis |
| Artificial Analysis Omniscience Accuracy | Knowledge | 55.3% | 82.3 | — | — | Artificial Analysis |
| HLE-Verified | Knowledge | 53.6% | — | — | — | Weiqi Zhai et al. |
| OSWorld 2.0 | Agentic | 47.9% | 78.8 | — | — | Mengqi Yuan et al. |
| Artificial Analysis Humanity's Last Exam | Knowledge | 47.9% | 76.5 | — | — | Artificial Analysis |
| Chess Puzzles | Reasoning | 47.0% | 84.3 | high effort | — | Epoch AI |
| GDPval-AA normalized | Agentic | 43.6% | 69.1 | — | — | Artificial Analysis |
| FrontierCode 1.1 Main | Coding | 43.6% | 72.3 | — | — | Cognition |
| BioMysteryBench Human Difficult | Knowledge | 43.5% | — | — | — | Anthropic |
| Artificial Analysis Intelligence Index | Knowledge | 39.1% | 71.2 | — | — | Artificial Analysis |
| Mystery Game Puzzles | Reasoning | 37.0% | 73.3 | high effort | — | Epoch AI |
| FrontierMath-Tier-4-v2-Private | Math | 36.6% | 65.5 | high effort | — | Epoch AI |
| Artificial Analysis Agentic Index | Agentic | 36.4% | 66.9 | — | — | Artificial Analysis |
| Code Migration | Coding | 34.8% | 67.2 | — | 21 Sept 2026 | Vals AI |
| AutomationBench | Agentic | 30.4% | 66.0 | — | — | Moonshot AI |
| Furniture Assembly | Reasoning | 26.7% | 58.5 | high effort | — | Epoch AI |
| Agents' Last Exam | Agentic | 26.3% | 66.0 | — | — | DeepSeek-AI |
| FrontierSWE v2 | Coding | 20.3% | 65.7 | — | — | Proximal |
| ApprenticeBench: end-to-end computer use, continual learning, and long-horizon agency on a real accounts-payable job | Agentic | 16.0% | 69.0 | — | — | NeoCognition |
| Terminal-Bench 3.0 | Agentic | 14.9% | 66.2 | — | — | Ryan Marten et al. |
| Critical Physics Tasks | Reasoning | 14.3% | 68.4 | — | — | Artificial Analysis |
| Terminal-Bench 4.0.0 | Agentic | 11.2% | 67.4 | high effort · mini-SWE-agent | 21 Sept 2026 | Terminal-Bench |
| Terminal-Bench 4.0 | Agentic | 6.1% | 63.7 | — | 21 Sept 2026 | Vals AI |
| Agent Arena task outcome | Agentic | 2.9 | 70.6 | high effort | 15 Sept 2026 | LMArena |
| ProgramBench | Coding | 0.0% | — | — | 21 Sept 2026 | Vals AI |
| Agent Arena steerability | Agentic | -0.3 | 67.0 | high effort | 15 Sept 2026 | LMArena |
| Agent Arena command recovery | Agentic | -2.1 | 65.0 | high effort | 15 Sept 2026 | LMArena |
51 benchmarks count, from 65 of 73 results. A grey row does not count. Too few models took that benchmark.