Model profile
Magistral Small
Evidence summary
Magistral Small has an estimated overall rank of #252; its 90% source-sensitivity interval is #133–#265. Its behavior-only rank is #245; company governance moves the combined estimate to #252. Published evidence spans 3 evals and 5 of 7 behavior components. Its strongest relative result is AA-Omniscience (hallucination_rate, #120 of 311); its weakest is PHARE (harm_resistance_diagnostic, #68 of 70).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗magistral-small
- Hugging Face ↗mistralai/Magistral-Small-2506
- Official model page ↗Exact model document · Reviewed official Hugging Face owner · official repository
- Release source ↗repository created at proxy
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 | #120 / 311 | 0.7548 | ↓ lower | Source ↗official | |
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #141 / 260 | 14.99 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #231 / 260 | 76.17 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #241 / 260 | 33.33 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #204 / 258 | 91.64 | ↑ higher | Source ↗official | |
| PHAREbias_resistance_diagnostic | #28 / 66 | 0.4802 | ↑ higher | Source ↗official | |
| PHAREhallucination_resistance_diagnostic | #64 / 70 | 0.6244 | ↑ higher | Source ↗official | |
| PHAREharm_resistance_diagnostic | #68 / 70 | 0.7623 | ↑ higher | Source ↗official | |
| PHAREjailbreak_resistance_diagnostic | #57 / 67 | 0.3709 | ↑ higher | Source ↗official |
