Model profile
Mimo V2.5
Evidence summary
Mimo V2.5 has an estimated overall rank of #97; its 90% source-sensitivity interval is #38–#214. Its behavior-only rank is #95; company governance moves the combined estimate to #97. Published evidence spans 7 evals and 5 of 7 behavior components. Its strongest relative result is AA-Omniscience (hallucination_rate, #25 of 327); its weakest is Inkling-Small model card — FORTRESS (harmful_refusal_rate, #9 of 10).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗mimo-v2-5-0424
- Hugging Face ↗XiaomiMiMo/MiMo-V2.5
- OpenRouter ↗xiaomi/mimo-v2.5
- 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 | #25 / 327 | ↓0.3193 | Source ↗official | |
| Arena Factuality — Text Arena (factuality-only weighting)factuality_bt_rating | #46 / 112 | ↑1441.0 | Source ↗official | |
| BullshitBench v2clear_pushback_rate | #51 / 106 | ↑0.295 | Source ↗official | |
| DystopiaBenchbasaglia_score | #7 / 50 | ↓30.8 | Source ↗official | |
| DystopiaBenchbaudrillard_score | #4 / 50 | ↓27.13 | Source ↗official | |
| DystopiaBenchhuxley_score | #10 / 50 | ↓33.8 | Source ↗official | |
| DystopiaBenchlaguardia_score | #7 / 50 | ↓39.97 | Source ↗official | |
| DystopiaBenchorwell_score | #9 / 50 | ↓30.8 | Source ↗official | |
| DystopiaBenchpetrov_score | #11 / 50 | ↓44.4 | Source ↗official | |
| Inkling-Small model card — FORTRESSbenign_answer_rate | #7 / 10 | ↑94.6 | Source ↗official | |
| Inkling-Small model card — FORTRESSharmful_refusal_rate | #9 / 10 | ↑64.8 | Source ↗official | |
| Inkling-Small model card — StrongREJECTsafety_rate | #3 / 10 | ↑99.3 | Source ↗official | |
| SM-Benchadversarial | #45 / 78 | ↑80.49 | Source ↗official | |
| SM-Benchambiguous_interpretation | #54 / 78 | ↑81.85 | Source ↗official | |
| SM-Benchanti_hallucination | #50 / 78 | ↑88.48 | Source ↗official | |
| SM-Bencheq_boundaries | #17 / 78 | ↑70.51 | Source ↗official | |
| SM-Benchoverfit | #36 / 78 | ↑75.96 | Source ↗official | |
| TACbase_welfare_rate | #55 / 74 | ↑21.79 | Source ↗self run |
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.