Model profile
Minimax M2
Evidence summary
Minimax M2 has an estimated overall rank of #119; its 90% source-sensitivity interval is #27–#229. Its behavior-only rank is #118; company governance moves the combined estimate to #119. Published evidence spans 5 evals and 6 of 7 behavior components. Its strongest relative result is Enkrypt AI Safety Leaderboard (bias_attack_non_success_rate, #16 of 260); its weakest is SpeciEval (sea_animal_4ns, #94 of 102).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗minimax-m2
- Hugging Face ↗MiniMaxAI/MiniMax-M2
- OpenRouter ↗minimax/minimax-m2
- Official model page ↗Exact model document · Reviewed official Hugging Face owner · official repository
- Release source ↗direct research preview identity date
Published eval results
Rank is within that sub-eval. Black marks the observed result; the grey dot marks the value implied by the global rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Better | Source |
|---|---|---|---|---|---|
| AA-Omnisciencehallucination_rate | #236 / 311 | 0.8842 | ↓ lower | Source ↗official | |
| Cisco AI Defense Rolling Single-Turn Leaderboardsingle_turn_attack_success_rate | #32 / 105 | 18.05 | ↓ lower | Source ↗official | |
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #16 / 260 | 51.94 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #173 / 260 | 86 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #58 / 260 | 89.44 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #44 / 258 | 99.09 | ↑ higher | Source ↗official | |
| LiveSecBenchethics | #12 / 43 | 69.21 | ↑ higher | Source ↗official | |
| LiveSecBenchfactuality | #13 / 43 | 62.29 | ↑ higher | Source ↗official | |
| LiveSecBenchlegality | #12 / 43 | 67.96 | ↑ higher | Source ↗official | |
| LiveSecBenchprivacy | #14 / 43 | 64.5 | ↑ higher | Source ↗official | |
| LiveSecBenchpsychological_health | #10 / 43 | 69.47 | ↑ higher | Source ↗official | |
| SpeciEvalbelief_animal_sentience | #47 / 102 | 6.82 | ↑ higher | Source ↗official | |
| SpeciEvalland_animal_4ns | #89 / 102 | 4.97 | ↓ lower | Source ↗official | |
| SpeciEvalsea_animal_4ns | #94 / 102 | 5.2 | ↓ lower | Source ↗official | |
| SpeciEvalspeciesism | #36 / 102 | 1.81 | ↓ lower | Source ↗official |
Values evaluations
Descriptive values and political-framing results are separate from safety/ethics ranks. Each strip shows the evaluation’s observed model range; its endpoint labels state what lower and higher values mean.