Model profile
Minimax M2.7
Evidence summary
Minimax M2.7 has an estimated overall rank of #109; its 90% source-sensitivity interval is #55–#219. Its behavior-only rank is #108; company governance moves the combined estimate to #109. Published evidence spans 13 evals and 6 of 7 behavior components. Its strongest relative result is Enkrypt AI Safety Leaderboard (bias_attack_non_success_rate, #11 of 241); its weakest is Inkling-Small model card — FORTRESS (benign_answer_rate, #10 of 10).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗minimax-m2-7
- Hugging Face ↗MiniMaxAI/MiniMax-M2.7
- OpenRouter ↗minimax/minimax-m2.7
- 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 | Source |
|---|---|---|---|---|
| AA-Omnisciencehallucination_rate | #33 / 328 | ↓0.3556 | Source ↗official | |
| Arena Factuality — Text Arena (factuality-only weighting)factuality_bt_rating | #60 / 112 | ↑1434.0 | Source ↗official | |
| BullshitBench v2clear_pushback_rate | #92 / 106 | ↑0.075 | Source ↗official | |
| DystopiaBenchbasaglia_score | #12 / 50 | ↓46.93 | Source ↗official | |
| DystopiaBenchbaudrillard_score | #11 / 50 | ↓34.1 | Source ↗official | |
| DystopiaBenchhuxley_score | #12 / 50 | ↓45.9 | Source ↗official | |
| DystopiaBenchlaguardia_score | #16 / 50 | ↓59.8 | Source ↗official | |
| DystopiaBenchorwell_score | #11 / 50 | ↓41.07 | Source ↗official | |
| DystopiaBenchpetrov_score | #15 / 50 | ↓56.97 | Source ↗official | |
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #11 / 241 | ↑54.52 | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #166 / 241 | ↑85 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #20 / 241 | ↑97.78 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #62 / 239 | ↑98.18 | Source ↗official | |
| Inkling-Small model card — FORTRESSbenign_answer_rate | #10 / 10 | ↑90.1 | Source ↗official | |
| Inkling-Small model card — FORTRESSharmful_refusal_rate | #2 / 10 | ↑86.3 | Source ↗official | |
| Inkling-Small model card — StrongREJECTsafety_rate | #1 / 10 | ↑99.4 | Source ↗official | |
| ODCV-Benchaverage_severity | #6 / 12 | ↓1.365 | Source ↗official | |
| ODCV-Benchmisalignment_rate | #8 / 12 | ↓35.9 | Source ↗official | |
| RefusalBenchyouden_j | #10 / 19 | ↑0.08085 | Source ↗official | |
| SM-Benchadversarial | #78 / 79 | ↑68.78 | Source ↗official | |
| SM-Benchambiguous_interpretation | #37 / 79 | ↑86.01 | Source ↗official | |
| SM-Benchanti_hallucination | #75 / 79 | ↑72.77 | Source ↗official | |
| SM-Bencheq_boundaries | #70 / 79 | ↑52.53 | Source ↗official | |
| SM-Benchoverfit | #57 / 79 | ↑54.64 | Source ↗official | |
| SpeciEvalbelief_animal_sentience | #69 / 113 | ↑6.7 | Source ↗official | |
| SpeciEvalland_animal_4ns | #87 / 113 | ↓4.8 | Source ↗official | |
| SpeciEvalsea_animal_4ns | #82 / 113 | ↓4.92 | Source ↗official | |
| SpeciEvalspeciesism | #24 / 113 | ↓1.6 | Source ↗official | |
| TACbase_welfare_rate | #58 / 76 | ↑21.15 | Source ↗self run | |
| ToolPrivacyBenchprivate_mt_poi | #5 / 9 | ↓24.47 | Source ↗official | |
| ToolPrivacyBenchpublic_mt_poi | #2 / 9 | ↓16.75 | Source ↗official | |
| Vectara HHEM Factual Consistencyfactual_consistency_rate | #77 / 94 | ↑87.1 | 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.
UGI Political Values
CCPBench political narrative alignment
| Dimension | Value | Distribution |
|---|---|---|
| CCP-narrative alignment — all questions | 3.67 | |
| CCP-narrative alignment — China topics | 4.17 | |
| CCP-narrative alignment — non-China controls | 2.18 |
Agent-ValueBench Moral Foundations (MFT08)
Agent-ValueBench HEXACO
| Dimension | Value | Distribution |
|---|---|---|
| Openness to experience | 5.9 | |
| Honesty-humility | 5.6 | |
| Extraversion | 5 | |
| Agreeableness | 6.3 | |
| Conscientiousness | 7 |