Model profile
Mistral Small 3
Evidence summary
Mistral Small 3 has an estimated overall rank of #239; its 90% source-sensitivity interval is #180–#264. Its behavior-only rank is #233; company governance moves the combined estimate to #239. Published evidence spans 8 evals and 6 of 7 behavior components. Its strongest relative result is Enkrypt AI Safety Leaderboard (cbrn_attack_non_success_rate, #83 of 260); its weakest is HELM Safety (harmbench, #76 of 80).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗mistral-small-3
- Hugging Face ↗mistralai/Mistral-Small-24B-Instruct-2501
- 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 |
|---|---|---|---|---|---|
| AIRBench 2024 Safety Scenariossafety_scenarios | #76 / 80 | 0.328 | ↑ higher | Source ↗official | |
| Confabulationsconfabulation_rate | #43 / 52 | 38.61 | ↓ lower | Source ↗official | |
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #137 / 260 | 15.25 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #83 / 260 | 90.67 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #207 / 260 | 45.56 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #124 / 258 | 96.55 | ↑ higher | Source ↗official | |
| HELM Safetyanthropic_red_team | #69 / 80 | 0.964 | ↑ higher | Source ↗official | |
| HELM Safetybbq | #42 / 80 | 0.933 | ↑ higher | Source ↗official | |
| HELM Safetyharmbench | #76 / 80 | 0.452 | ↑ higher | Source ↗official | |
| HELM Safetysimple_safety_tests | #71 / 80 | 0.933 | ↑ higher | Source ↗official | |
| HELM Safetyxstest | #48 / 80 | 0.953 | ↑ higher | 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.