MiniMax
availableShows if the model has enough results for an index.MiniMax M2.7
MiniMax M2.7 is a non-reasoning model from MiniMax. 30 benchmarks count toward its score, in 6 categories.
IndexOverall score out of 100.52.9 ±4.5
CoverageShare of the index weight with results.85%
SpeedOutput tokens per second.56/s
Input / 1MUS dollars per 1M input tokens.$0.3
Output / 1MUS dollars per 1M output tokens.$1.2
ContextMaximum tokens in one request.205K
EloLMArena rating and rank.1405 (#138)
The index is a score out of 100. The ± range shows how much it can change.
69,169 votes. Elo shows what people prefer. It does not change the score.
CapabilitiesScore per category, out of 100.
Out of 100Results
30 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 | 91.0% | 56.2 | — | 16 Apr 2026 | Vals AI |
| Artificial Analysis GPQA Diamond | Knowledge | 87.4% | 58.1 | — | — | Artificial Analysis |
| GPQA Diamond | Knowledge | 87.0% | 58.6 | — | — | David Rein et al. |
| GPQA Diamond | Knowledge | 86.6% | 58.2 | — | 1 Sept 2026 | Vals AI |
| τ²-Bench Tool-Agent-User Evaluation | Agentic | 84.8% | 59.7 | — | — | Victor Barres et al. |
| MMLU-Pro first-party comparison snapshot | Knowledge | 80.8% | 47.8 | — | — | Arcee AI |
| MMLU Pro | Knowledge | 80.4% | 47.2 | — | 1 Sept 2026 | Vals AI |
| AIME25 first-party comparison snapshot | Math | 80.0% | — | — | — | Arcee AI |
| LiveCodeBench | Coding | 79.9% | 57.0 | — | 1 Sept 2026 | Vals AI |
| Artificial Analysis Long Context Reasoning | Reasoning | 78.3% | 62.4 | — | — | Artificial Analysis |
| Artificial Analysis IFBench | Instruction | 75.7% | 65.9 | — | — | Artificial Analysis |
| SWE-bench Verified (mini-swe-agent-v2) | Coding | 75.4% | 58.0 | — | — | Arcee AI |
| SWE-bench | Coding | 73.8% | 56.7 | — | 1 Sept 2026 | Vals AI |
| React Native Evals | Coding | 71.4% | 47.6 | — | — | Callstack |
| MLE-Bench Lite | Agentic | 66.6% | — | — | — | MiniMax |
| MM-ClawBench | Agentic | 62.7% | — | — | — | MiniMax |
| SWE-bench Pro | Coding | 56.2% | 58.3 | — | — | Xiang Deng et al. |
| VIBE-Pro | Coding | 55.6% | — | — | — | MiniMax |
| Multi-SWE Bench | Coding | 52.7% | — | — | — | MiniMax |
| Artificial Analysis Coding Index | Coding | 52.6% | 56.0 | — | — | Artificial Analysis |
| SWE-Rebench | Coding | 51.9% | — | — | — | Nebius |
| Artificial Analysis SciCode | Coding | 50.1% | 62.4 | — | — | Artificial Analysis |
| Claw-Eval | Agentic | 48.7% | 35.0 | — | — | Bowen Ye et al. |
| Terminal-Bench 2.1 | Agentic | 48.7% | 52.8 | — | 21 Sept 2026 | Vals AI |
| Terminal-Bench 2.0 | Agentic | 47.2% | 56.4 | — | 4 Jun 2026 | Vals AI |
| Toolathlon | Agentic | 46.3% | 61.1 | — | — | OpenAI |
| Gert Labs Composite Game Benchmark | Agentic | 40.4% | 48.4 | — | — | Gert Labs |
| NL2Repo | Coding | 39.8% | 56.1 | — | — | MiniMax |
| Artificial Analysis Humanity's Last Exam | Knowledge | 29.6% | 56.7 | — | — | Artificial Analysis |
| Artificial Analysis Omniscience Accuracy | Knowledge | 26.8% | 47.0 | — | — | Artificial Analysis |
| GDPval-AA normalized | Agentic | 24.9% | 54.7 | — | — | Artificial Analysis |
| Artificial Analysis Intelligence Index | Knowledge | 22.8% | 50.9 | — | — | Artificial Analysis |
| Artificial Analysis Agentic Index | Agentic | 16.8% | 50.9 | — | — | Artificial Analysis |
| Vibe Code Bench v1.1 | Coding | 11.9% | 47.1 | OpenHands | 21 Sept 2026 | Vals AI |
| APEX-Agents-AA | Agentic | 10.6% | 49.5 | — | — | Artificial Analysis / Mercor |
| Code Migration | Coding | 8.7% | 50.5 | — | 21 Sept 2026 | Vals AI |
| IOI v1 | Coding | 4.9% | 42.7 | — | 9 Aug 2026 | Vals AI |
| Critical Physics Tasks | Reasoning | 0.6% | 39.6 | — | — | Artificial Analysis |
| ProgramBench | Coding | 0.0% | — | — | 21 Sept 2026 | Vals AI |
| Agent Arena steerability | Agentic | -6.6 | 59.9 | — | 15 Sept 2026 | LMArena |
| Agent Arena task outcome | Agentic | -18.0 | 47.0 | — | 15 Sept 2026 | LMArena |
| Agent Arena command recovery | Agentic | -26.3 | 37.7 | — | 15 Sept 2026 | LMArena |
30 benchmarks count, from 35 of 42 results. A grey row does not count. Too few models took that benchmark.