Google
availableShows if the model has enough results for an index.Gemini 3 Pro
Gemini 3 Pro is a non-reasoning model from Google. 32 benchmarks count toward its score, in 8 categories.
IndexOverall score out of 100.60.5 ±2.9
CoverageShare of the index weight with results.100%
SpeedOutput tokens per second.109/s
Input / 1MUS dollars per 1M input tokens.$2
Output / 1MUS dollars per 1M output tokens.$12
ContextMaximum tokens in one request.2M
EloLMArena rating and rank.1479 (#17)
The index is a score out of 100. The ± range shows how much it can change.
40,654 votes. Elo shows what people prefer. It does not change the score.
CapabilitiesScore per category, out of 100.
Out of 100Results
32 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. |
|---|---|---|---|---|---|---|
| AIME | Math | 96.7% | 58.8 | — | 16 Apr 2026 | Vals AI |
| MATH 500 | Math | 96.4% | 49.5 | — | 9 Jan 2026 | Vals AI |
| MGSM | Multilingual | 93.9% | — | — | 9 Jan 2026 | Vals AI |
| GPQA diamond | Knowledge | 92.6% | 63.7 | — | — | Epoch AI |
| Artificial Analysis Global-MMLU-Lite | Multilingual | 92.2% | — | — | — | Artificial Analysis |
| Artificial Analysis LiveCodeBench | Coding | 91.7% | — | — | — | Artificial Analysis |
| GPQA Diamond | Knowledge | 91.7% | 62.9 | — | 1 Sept 2026 | Vals AI |
| OTIS Mock AIME 2024-2025 | Math | 91.4% | 62.6 | — | — | Epoch AI |
| τ²-bench Telecom | Agentic | 91.0% | 64.1 | high effort · Sierra | 2 Mar 2026 | Sierra Research |
| Artificial Analysis GPQA Diamond | Knowledge | 90.8% | 61.6 | — | — | Artificial Analysis |
| MMLU Pro | Knowledge | 90.1% | 62.5 | — | 1 Sept 2026 | Vals AI |
| Artificial Analysis MMLU-Pro | Knowledge | 89.8% | — | — | — | Artificial Analysis |
| V* | Multimodal | 88.0% | 46.7 | — | — | Z.AI |
| VideoMMMU | Multimodal | 87.6% | — | — | — | Qwen |
| MMMU Pro | Multimodal | 87.5% | 69.8 | — | 1 Sept 2026 | Vals AI |
| τ²-Bench Tool-Agent-User Evaluation | Agentic | 87.1% | 61.3 | — | — | Victor Barres et al. |
| MathVision | Multimodal | 86.6% | — | — | — | Qwen |
| LiveCodeBench | Coding | 86.4% | 62.9 | — | 1 Sept 2026 | Vals AI |
| CharXiv Reasoning | Multimodal | 81.4% | 58.8 | — | — | CharXiv authors |
| Massive Multi-discipline Multimodal Understanding Pro | Multimodal | 81.0% | 59.2 | — | — | MMMU-Pro authors |
| τ²-bench Airline | Agentic | 80.5% | 56.6 | high effort · Sierra | 2 Mar 2026 | Sierra Research |
| Artificial Analysis MMMU-Pro | Multimodal | 80.2% | 65.1 | — | — | Artificial Analysis |
| SWE-bench | Coding | 76.4% | 58.8 | — | 1 Sept 2026 | Vals AI |
| Artificial Analysis Long Context Reasoning | Reasoning | 76.0% | 60.8 | — | — | Artificial Analysis |
| τ²-bench Retail | Agentic | 75.9% | 53.3 | high effort · Sierra | 30 Apr 2026 | Sierra Research |
| ARC-AGI-1 (semi-private) | Reasoning | 75.0% | 62.6 | — | — | ARC Prize Foundation |
| SWE-bench Verified | Coding | 74.2% | 57.1 | mini-SWE-agent | 26 Feb 2026 | SWE-bench team |
| SWE-Bench verified | Coding | 72.9% | 56.0 | — | — | Epoch AI |
| ScreenSpot Pro | Multimodal | 72.7% | 54.3 | — | — | Kaixin Li et al. |
| Artificial Analysis IFBench | Instruction | 70.4% | 60.5 | — | — | Artificial Analysis |
| SWE-bench Verified | Coding | 69.6% | 53.4 | high effort · mini-SWE-agent | 1 Sept 2026 | SWE-bench team |
| SWE-bench Multilingual | Multilingual | 68.7% | — | mini-SWE-agent | 20 Feb 2026 | SWE-bench team |
| SWE-bench Multilingual | Coding | 68.7% | — | mini-SWE-agent | 2 Sept 2026 | SWE-bench team |
| EuroEval Polish | Multilingual | 64.8% | 83.2 | — | — | EuroEval |
| Gert Labs Composite Game Benchmark | Agentic | 63.2% | 68.6 | — | — | Gert Labs |
| Artificial Analysis Omniscience Accuracy | Knowledge | 55.8% | 82.9 | — | — | Artificial Analysis |
| Terminal-Bench 2.0 | Agentic | 55.1% | 62.0 | — | 4 Jun 2026 | Vals AI |
| Terminal-Bench 1.0 | Agentic | 51.3% | 55.0 | — | 12 Jan 2026 | Vals AI |
| Artificial Analysis Humanity's Last Exam | Knowledge | 39.7% | 67.6 | — | — | Artificial Analysis |
| IOI v1 | Coding | 38.8% | 62.0 | — | 9 Aug 2026 | Vals AI |
| FrontierMath-2025-02-28-Private | Math | 37.6% | 65.7 | — | — | Epoch AI |
| ARC-AGI-2 (semi-private) | Reasoning | 31.1% | 55.4 | — | — | ARC Prize Foundation |
| Chess Puzzles | Reasoning | 31.0% | 63.6 | — | — | Epoch AI |
| Artificial Analysis Intelligence Index | Knowledge | 28.0% | 57.4 | — | — | Artificial Analysis |
| FrontierMath-Tier-4-2025-07-01-Private | Math | 18.8% | 63.8 | — | — | Epoch AI |
| τ²-bench Banking | Agentic | 18.0% | 11.8 | high effort · Sierra | 4 Aug 2026 | Sierra Research |
| Vibe Code Bench v1.1 | Coding | 14.3% | 48.1 | OpenHands | 21 Sept 2026 | Vals AI |
| JobBench | Agentic | 11.4% | 42.0 | — | — | Yuetai Li et al. |
| Critical Physics Tasks | Reasoning | 9.1% | 57.5 | — | — | Artificial Analysis |
32 benchmarks count, from 41 of 49 results. A grey row does not count. Too few models took that benchmark.