MiMo-V2-Flash.
Open weights
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MMLU-Pro | 84.9% | #32 of 167 | Lab model card | |
| LongBench v2 | 60.6% | #13 of 35 | Lab model card | |
| LMArena (Elo) | 1,392 | #140 of 335 | Official leaderboard | |
| Humanity's Last Exam | 22.1% | #99 of 154 | Lab model card | |
| Longform Creative Writing (EQ-Bench) | 53.1 | #80 of 119 | Official leaderboard | |
| BullshitBench | 16.4% | #77 of 107 | Official leaderboard | |
| MRCR v2 | 45.7% | #23 of 30 | Lab model card | |
| BullshitBench V2 | 16% | #101 of 129 | Official leaderboard | |
| Arena-Hard v2.0 | 54.1% | #20 of 24 | Lab model card | |
| DesignArena ASCII Art | 1,012 | #75 of 77 | Official leaderboard |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| AIME 2025 | 94.1% | #26 of 143 | Lab model card | |
| GPQA Diamond | 84.3% | #82 of 302 | Lab model card | |
| HMMT 2025 | 84.4% | #31 of 54 | Lab model card |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| LiveCodeBench v6 | 85.1% | #19 of 90 | Lab model card | |
| SWE-bench Verified | 73.4% | #56 of 155 | Lab model card | |
| SWE-bench Multilingual | 71.7% | #40 of 73 | Lab model card | |
| DesignArena Website | 1,160 | #96 of 151 | Official leaderboard | |
| DesignArena Code | 1,146 | #96 of 143 | Official leaderboard | |
| DesignArena UI Component | 1,135 | #96 of 139 | Official leaderboard | |
| DesignArena 3D Design | 1,098 | #95 of 136 | Official leaderboard | |
| DesignArena Game Dev | 1,121 | #100 of 143 | Official leaderboard | |
| DesignArena Data Visualization | 1,135 | #95 of 118 | Official leaderboard | |
| WebDev Arena | 1,331 | #87 of 105 | arena.ai, via Epoch AI | |
| Code Arena | 1,331 | #107 of 128 | Official leaderboard | |
| DesignArena SVG | 1,020 | #88 of 103 | Official leaderboard | |
| ALE-Bench | 1,000 | #110 of 126 | Official leaderboard | |
| CF-Div2-Stepfun | 46.9% | #8 of 8 | Official leaderboard |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Terminal-Bench Hard | 30.5% | #4 of 13 | Lab model card | |
| tau2-bench | 80.3% | #34 of 79 | Lab model card | |
| PinchBench | 67.5% | #25 of 48 | Official leaderboard | |
| SnakeBench | 23 | #109 of 188 | Official leaderboard | |
| BrowseComp | 58.3% | #59 of 81 | Lab model card | |
| Pencil Puzzle Bench | 0% | #43 of 53 | Official leaderboard | |
| Terminal-Bench 2.0 | 38.5% | #53 of 63 | Lab model card | |
| LisanBench | 104.3 | #113 of 133 | Official leaderboard | |
| WildClawBench | 30.8% | #28 of 31 | Official leaderboard |
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| LiveSecBench | 57.2 | #14 of 34 | Official leaderboard |
The Self-Reported Index fits one score to the 15 capability results Xiaomi Corp reported for this model, on the scale of the Epoch Capabilities Index. Labs choose which benchmarks to show and how to run them, so read it as the lab's claim, not an independent measurement. The 90% interval runs from 140.7 to 145.4. How we compute it.
The benchmarks the lab chose to headline, with the scores it gave.
Every result in the longer documents the lab published with it.