Google
availableShows if the model has enough results for an index.Gemini 3 Flash
Gemini 3 Flash is a non-reasoning model from Google. 22 benchmarks count toward its score, in 7 categories.
IndexOverall score. 50 is the middle.55.3 ±3.9
CoverageShare of the index weight with results.95%
SpeedOutput tokens per second.213/s
Input / 1MUS dollars per 1M input tokens.$0.5
Output / 1MUS dollars per 1M output tokens.$3
ContextMaximum tokens in one request.1M
EloLMArena rating and rank.1467 (#30)
50 is the middle of the board. The range shows the doubt in the index.
30,225 votes. Elo shows what people prefer. It does not change the score.
CapabilitiesScore per category. 50 is the middle.
50 is the middleResults
22 counted| BenchmarkThe test name. | CategoryThe capability that the test measures. | ResultThe score from the publisher. | IndexThis result on the index scale. | RunThe settings of the run. | DateDate of the result. | Published byThe source of the result. |
|---|---|---|---|---|---|---|
| OTIS Mock AIME 2024-2025 | Math | 95.6% | 65.0 | high effort | — | Epoch AI |
| Artificial Analysis Global-MMLU-Lite | Multilingual | 92.7% | — | — | — | Artificial Analysis |
| τ²-bench Telecom | Agentic | 91.2% | 64.3 | high effort · Sierra | 2 Mar 2026 | Sierra Research |
| GPQA diamond | Knowledge | 89.4% | 60.8 | high effort | — | Epoch AI |
| τ²-bench Airline | Agentic | 82.5% | 58.0 | high effort · Sierra | 2 Mar 2026 | Sierra Research |
| Artificial Analysis GPQA Diamond | Knowledge | 81.2% | 51.8 | — | — | Artificial Analysis |
| Artificial Analysis MMMU-Pro | Multimodal | 78.6% | 63.1 | — | — | Artificial Analysis |
| τ²-bench Retail | Agentic | 76.8% | 53.9 | high effort · Sierra | 30 Apr 2026 | Sierra Research |
| SWE-Bench verified | Coding | 75.4% | 58.0 | — | — | Epoch AI |
| SWE-bench Multilingual | Multilingual | 72.7% | — | mini-SWE-agent | 20 Feb 2026 | SWE-bench team |
| SWE-bench Multilingual | Coding | 72.7% | — | mini-SWE-agent | 2 Sept 2026 | SWE-bench team |
| SimpleQA Verified | Knowledge | 66.8% | 83.4 | high effort | — | Epoch AI |
| Gert Labs Composite Game Benchmark | Agentic | 56.6% | 62.7 | — | — | Gert Labs |
| Artificial Analysis Long Context Reasoning | Reasoning | 55.3% | 46.5 | — | — | Artificial Analysis |
| Artificial Analysis IFBench | Instruction | 55.1% | 45.0 | — | — | Artificial Analysis |
| FrontierMath-Tiers-1-3-v2-Private | Math | 51.2% | 60.3 | — | — | Epoch AI |
| Claw-Eval | Agentic | 49.2% | 35.8 | — | — | Bowen Ye et al. |
| Artificial Analysis Omniscience Accuracy | Knowledge | 45.8% | 70.5 | — | — | Artificial Analysis |
| τ²-Bench Tool-Agent-User Evaluation | Agentic | 43.3% | 29.9 | — | — | Victor Barres et al. |
| Chess Puzzles | Reasoning | 40.0% | 75.2 | high effort | — | Epoch AI |
| FrontierMath-2025-02-28-Private | Math | 35.6% | 63.9 | — | — | Epoch AI |
| τ²-bench Banking | Agentic | 27.3% | 18.4 | high effort · Sierra | 4 Aug 2026 | Sierra Research |
| Mystery Game Puzzles | Reasoning | 20.0% | 55.2 | high effort | — | Epoch AI |
| Artificial Analysis Intelligence Index | Knowledge | 17.9% | 44.8 | — | — | Artificial Analysis |
| FrontierMath-Tier-4-v2-Private | Math | 17.1% | 56.1 | — | — | Epoch AI |
| Artificial Analysis Humanity's Last Exam | Knowledge | 15.0% | 40.8 | — | — | Artificial Analysis |
| JobBench | Agentic | 11.4% | 42.0 | — | — | Yuetai Li et al. |
| FrontierMath-Tier-4-2025-07-01-Private | Math | 4.2% | 48.0 | — | — | Epoch AI |
| Critical Physics Tasks | Reasoning | 1.4% | 41.3 | — | — | Artificial Analysis |
22 benchmarks count, from 26 of 29 results. A grey row does not count. Too few models took that benchmark.