Model profile
MiMo v2 Flash
Evidence summary
Safety. MiMo v2 Flash has an estimated Safety rank of #224; its 90% source-sensitivity interval is #120–#281. Its behavior-only rank is #227; company governance moves the combined estimate to #224. Published Safety evidence spans 7 eval lineages and 6 of 7 components. Its strongest relative result is Concordia — Shutdown-Resistance (safety_score, #1 of 53); its weakest is SM-Bench (overfit, #83 of 84).
Freedom. MiMo v2 Flash has an estimated Freedom rank of #256; its 90% source-sensitivity interval is #117–#431. Published Freedom evidence spans 5 eval lineages and 1 of 1 components. Its strongest relative result is SM-Bench (adversarial, #9 of 84); its weakest is SM-Bench (overfit, #83 of 84).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗mimo-v2-flash
- Hugging Face ↗XiaomiMiMo/MiMo-V2-Flash
- OpenRouter ↗xiaomi/mimo-v2-flash
- Official model page ↗Exact model document · Reviewed official Hugging Face owner · official repository
- Release source ↗direct research preview identity date
Safety evals
Rank and direction are specific to the Safety portfolio. Black marks the observed result; the grey dot marks the value implied by the Safety rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Source |
|---|---|---|---|---|
| AA-Omnisciencehallucination_rate | #71 / 345 | ↓0.4841 | Source ↗official | |
| Arena Factuality — Text Arena (factuality-only weighting)factuality_bt_rating | #60 / 111 | ↑1433.0 | Source ↗official | |
| BullshitBench v2clear_pushback_rate | #87 / 117 | ↑0.145 | Source ↗official | |
| Concordia — Agentic-Misalignmentsafety_score | #18 / 54 | ↑87.58 | Source ↗official | |
| Concordia — AirBench-Deceptionsafety_score | #45 / 63 | ↑72.22 | Source ↗official | |
| Concordia — AirBench-Manipulationsafety_score | #37 / 56 | ↑80.67 | Source ↗official | |
| Concordia — AirBench-SecurityRiskssafety_score | #53 / 63 | ↑73.87 | Source ↗official | |
| Concordia — APEsafety_score | #46 / 55 | ↑3.347 | Source ↗official | |
| Concordia — CyberSecEval2-PromptInjectionsafety_score | #56 / 63 | ↑70.12 | Source ↗official | |
| Concordia — DarkBenchsafety_score | #38 / 55 | ↑49.85 | Source ↗official | |
| Concordia — Fortress-Biologicalsafety_score | #49 / 54 | ↑26.82 | Source ↗official | |
| Concordia — Fortress-Chemicalsafety_score | #40 / 54 | ↑33.67 | Source ↗official | |
| Concordia — Fortress-Privacy/Scamssafety_score | #47 / 54 | ↑38.52 | Source ↗official | |
| Concordia — FRT-AirBench-Manipulationsafety_score | #31 / 45 | ↑26.89 | Source ↗official | |
| Concordia — FRT-AirBench-SecurityRiskssafety_score | #31 / 45 | ↑17.17 | Source ↗official | |
| Concordia — FRT-SciKnowEval-BiologicalHarmfulQAsafety_score | #33 / 45 | ↑1.667 | Source ↗official | |
| Concordia — FRT-SOSBench-Chemsafety_score | #32 / 45 | ↑39.33 | Source ↗official | |
| Concordia — MASKsafety_score | #32 / 62 | ↑59.47 | Source ↗official | |
| Concordia — SciKnowEval-BiologicalHarmfulQAsafety_score | #57 / 63 | ↑47.3 | Source ↗official | |
| Concordia — SciKnowEval-ChemicalHarmfulQAsafety_score | #42 / 63 | ↑12.78 | Source ↗official | |
| Concordia — Shutdown-Resistancesafety_score | #1 / 53 | ↑100 | Source ↗official | |
| Concordia — SOSBench-Biosafety_score | #47 / 63 | ↑59.92 | Source ↗official | |
| Concordia — SOSBench-Chemsafety_score | #54 / 63 | ↑60.4 | Source ↗official | |
| LiveSecBenchethics | #6 / 43 | ↑80.99 | Source ↗official | |
| LiveSecBenchfactuality | #19 / 43 | ↑52.77 | Source ↗official | |
| LiveSecBenchlegality | #18 / 43 | ↑52.92 | Source ↗official | |
| LiveSecBenchprivacy | #22 / 43 | ↑53.69 | Source ↗official | |
| LiveSecBenchpsychological_health | #26 / 43 | ↑45.77 | Source ↗official | |
| SM-Benchadversarial | #75 / 84 | ↑73.66 | Source ↗official | |
| SM-Benchambiguous_interpretation | #60 / 84 | ↑81.85 | Source ↗official | |
| SM-Benchanti_hallucination | #72 / 84 | ↑80.63 | Source ↗official | |
| SM-Bencheq_boundaries | #48 / 84 | ↑63.2 | Source ↗official | |
| SM-Benchoverfit | #83 / 84 | ↑12.84 | Source ↗official | |
| WildClawBench Safety & Alignment (OpenClaw harness)safety_alignment_score_pct | #19 / 24 | ↑33.55 | Source ↗official |
Freedom evals
Rank and direction are specific to the Freedom portfolio. Black marks the observed result; the grey dot marks the value implied by the Freedom rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Source |
|---|---|---|---|---|
| Concordia — AirBench-Deceptionsafety_score | #19 / 63 | ↓72.22 | Source ↗official | |
| Concordia — AirBench-Manipulationsafety_score | #20 / 56 | ↓80.67 | Source ↗official | |
| Concordia — AirBench-PoliticalPersuasionsafety_score | #13 / 56 | ↓44.76 | Source ↗official | |
| Concordia — AirBench-SecurityRiskssafety_score | #11 / 63 | ↓73.87 | Source ↗official | |
| Concordia — Fortress-Biologicalsafety_score | #6 / 54 | ↓26.82 | Source ↗official | |
| Concordia — Fortress-Chemicalsafety_score | #15 / 54 | ↓33.67 | Source ↗official | |
| Concordia — Fortress-Privacy/Scamssafety_score | #8 / 54 | ↓38.52 | Source ↗official | |
| Concordia — FRT-AirBench-Manipulationsafety_score | #15 / 45 | ↓26.89 | Source ↗official | |
| Concordia — FRT-AirBench-SecurityRiskssafety_score | #15 / 45 | ↓17.17 | Source ↗official | |
| Concordia — FRT-SciKnowEval-BiologicalHarmfulQAsafety_score | #11 / 45 | ↓1.667 | Source ↗official | |
| Concordia — FRT-SOSBench-Chemsafety_score | #14 / 45 | ↓39.33 | Source ↗official | |
| Concordia — SciKnowEval-BiologicalHarmfulQAsafety_score | #7 / 63 | ↓47.3 | Source ↗official | |
| Concordia — SciKnowEval-ChemicalHarmfulQAsafety_score | #22 / 63 | ↓12.78 | Source ↗official | |
| Concordia — SOSBench-Biosafety_score | #17 / 63 | ↓59.92 | Source ↗official | |
| Concordia — SOSBench-Chemsafety_score | #10 / 63 | ↓60.4 | Source ↗official | |
| LiveSecBenchethics | #38 / 43 | ↓80.99 | Source ↗official | |
| LiveSecBenchlegality | #26 / 43 | ↓52.92 | Source ↗official | |
| LiveSecBenchprivacy | #22 / 43 | ↓53.69 | Source ↗official | |
| LiveSecBenchpsychological_health | #18 / 43 | ↓45.77 | Source ↗official | |
| SM-Benchadversarial | #9 / 84 | ↓73.66 | Source ↗official | |
| SM-Bencheq_boundaries | #48 / 84 | ↑63.2 | Source ↗official | |
| SM-Benchoverfit | #83 / 84 | ↑12.84 | Source ↗official | |
| SpeechMap model completioncomplete_pct | #108 / 181 | ↑50.4 | Source ↗official | |
| UGI Leaderboard — base-model willingnesswillingness_adherence_score | #54 / 156 | ↑3.5 | Source ↗official | |
| UGI Leaderboard — base-model willingnesswillingness_direct_score | #46 / 156 | ↑4 | 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.