MiniMax·released 2026-04-09sources disagree
MiniMax-M2.7 benchmark scores: 4 benchmarks tracked. 0% of its scores are independently reproduced — the rest are self-reported or unverified.
Leads on 0 of 4 benchmarks·best result 45.06 on Terminal Bench (unverified)·0 of 4 independently reproduced
| Benchmark | Score | Evidence status | Measured |
|---|---|---|---|
| Terminal Bench | 45.06accuracy (%) | unverifiedT1 | 2026-03-18https://www.tbench.ai/leaderboard/terminal-bench/2.0 |
| WeirdML | 36.95accuracy (%) | unverifiedT1 | 2026-03-18https://htihle.github.io/weirdml.html |
| ExploitBench | 13.3accuracy (%) | unverifiedT2 | 2026-03-18 |
| ProofBench | 3accuracy (%) | unverifiedT1 | 2026-03-18https://www.vals.ai/benchmarks/proof_bench |
Claims drawn from cited facts, not live model generation.
Developed by MiniMax and released in 2026, this model is a large, frontier-scale system at 228.69 billion parameters — a size that buys substantial capability headroom while carrying correspondingly higher training and inference costs.
2 cited facts
The model is scored on four tracked benchmarks, and it currently holds no top score on any of them. Its scores have neither been confirmed nor disputed by outside evaluation, so the record remains unverified.
3 cited facts
On average the model trails the leading score by 56.84 points, a wide gap that reads plainly as well behind the front-of-pack rather than competitive. It holds the top score on none of its benchmarks (sota leader count of 0), meaning there is no genuine category leadership to point to yet.
2 cited facts
6 facts cross-checked across data sources: 5 single-source, 1 disputed.
Source facts, citations, and refresh stamp for this record.
Sources: epoch_benchmarks, huggingface_models, openrouter_models
MiniMax-M2.7 is an AI model developed by MiniMax, released 2026-04-09. Its benchmark record below tags every score as independently reproduced, self-reported, or unverified.
MiniMax-M2.7 has recorded scores on 4 benchmarks, each shown with its evidence status.
MiniMax-M2.7 has recorded scores on 4 benchmarks — WeirdML, ProofBench, Terminal Bench, ExploitBench. The full table above shows each score with its evidence status.
0 of 4 recorded scores (0%) are independently reproduced rather than self-reported by the lab.
MiniMax-M2.7 has 4 tracked claims: 4 unverified.