Model profile
Mistral 7B
Evidence summary
Safety. Mistral 7B has an estimated Safety rank of #301; its 90% source-sensitivity interval is #153–#327. Its behavior-only rank is #293; company governance moves the combined estimate to #301. Published Safety evidence spans 4 eval lineages and 4 of 7 components. Its strongest relative result is Enkrypt AI Safety Leaderboard (bias_attack_non_success_rate, #3 of 248); its weakest is HalluVerse-M3 Hallucination Recognition (hallucination_recognition_accuracy, #14 of 14).
Freedom. Mistral 7B has an estimated Freedom rank of #29; its 90% source-sensitivity interval is #39–#253. Published Freedom evidence spans 3 eval lineages and 1 of 1 components. Its strongest relative result is Enkrypt AI Safety Leaderboard (toxicity_attack_non_success_rate, #17 of 246); its weakest is Enkrypt AI Safety Leaderboard (cbrn_attack_non_success_rate, #157 of 248).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Official model documentation ↗Family-level model document · mistral · first party
- Release source ↗curated outward evidence
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 |
|---|---|---|---|---|
| COMPL-AI AI-Identity Disclosurescore | #7 / 14 | ↑0.9452 | Source ↗official | |
| COMPL-AI LLM RuLES Multi-Turn Rule Followingscore | #11 / 14 | ↑0.2931 | Source ↗official | |
| COMPL-AI TensorTrust Goal-Hijacking Resistancescore | #4 / 13 | ↑0.5155 | Source ↗official | |
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #3 / 248 | ↑86.56 | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #90 / 248 | ↑89.83 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #224 / 248 | ↑36.67 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #230 / 246 | ↑78.18 | Source ↗official | |
| HalluVerse-M3 Hallucination Recognitionhallucination_recognition_accuracy | #14 / 14 | ↑56.06 | Source ↗official | |
| MedSafetyBenchmedical_safety_score | #27 / 30 | ↑53.19 | 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 |
|---|---|---|---|---|
| COMPL-AI LLM RuLES Multi-Turn Rule Followingscore | #4 / 14 | ↓0.2931 | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #157 / 248 | ↓89.83 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #25 / 248 | ↓36.67 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #17 / 246 | ↓78.18 | Source ↗official | |
| MedSafetyBenchmedical_safety_score | #4 / 30 | ↓53.19 | Source ↗official |