AI Safety Evidence

← All models

Developer

Mistral AI

39 indexed models; 21 currently meet the evidence threshold for the overall ranking. Together they have results from 58 evaluations.

Official Mistral AI model information ↗

Company governance evidence

Mistral AI is represented at -1.20 SD relative to the matched Future of Life Institute edition. Provenance: Direct. This company-level evidence contributes 10% of overall model rank.

Models by Mistral AI

ModelEvalsComponentsRankRelease date
Mistral Large66/7732024-02-26
Mistral Medium45/71472023-12-11
Mistral Small 455/71582026-01-23
Magistral Medium66/71632025-06-10
Mistral Small 3.142/71772025-03-11
Mistral Small 3.235/72312025-06-19
Mistral 7B34/7243—
Mistral Medium 343/72442025-05-07
Mistral 7B Instruct227/72452023-09-27
Mistral Medium 3.544/72522026-03-31
Mistral Small 396/72542025-01-30
Mistral Nemo55/72622024-07-18
Mistral Medium 3.176/72692025-08-12
Mixtral 8X22B Instruct95/72752024-04-17
Mistral Large 3146/72872025-12-01
Magistral Small35/72892025-06-04
Mistral Large 2125/72932024-07-24
Mixtral 8X7B Instruct116/73042023-12-10
Magistral Small 250934/73062025-09-12
Ministral 3 14B32/73082025-10-31
Ministral 3 8B42/73092025-10-31
Devstral11/7——
Devstral 222/7——
Devstral Small11/7——
Devstral Small 211/7——
Devstral Small 250511/7—2025-05-12
Ministral 3 3B22/7—2025-10-31
Ministral 3B11/7——
Ministral 3B 251211/7——
Ministral 8B 24.1024/7—2024-10-16
Ministral 8B 24.10 Moderated14/7—2024-10-16
Mistral Large 2402 Moderated14/7—2024-02-26
Mistral Large Latest13/7——
Mistral Medium 250811/7——
Mistral Nemo Minitron 8B Instruct13/7——
Mistral Small23/7——
Mistral Small 211/7——
Mistral Small Latest13/7——
Voxtral Small 24B 250722/7——
Evaluations covering Mistral AI models (58)

AA-Omniscience · AbstentionBench · Adversarial Robustness · AgentDrive Safety Compliance · AgentHarm · AILuminate General Purpose AI Chat · AIRBench 2024 Safety Scenarios · Alignment Leaderboard · AnimalHarmBench · Arena Factuality — Text Arena (factuality-only weighting) · AuAu Authoritarian Response Audit · BlueBench AttaQ-100 · BullshitBench v2 · CASE-Bench · ChineseSafe · Cisco AI Defense Rolling Single-Turn Leaderboard · COMPL-AI AI-Identity Disclosure · COMPL-AI LLM RuLES Multi-Turn Rule Following · COMPL-AI TensorTrust Goal-Hijacking Resistance · Confabulations · Contextual MoralChoice · DSPSafeBench · DystopiaBench · Emergent Collusion · Enkrypt AI Safety Leaderboard · FlagEval Safety and Values · FORTRESS · HalluVerse-M3 Hallucination Recognition · HarmBench · HELM Safety · HUMAINE Trust, Ethics and Safety · JailBench · Large-scale Moral Machine experiment on LLMs · LiveSecBench · MANTA · MASK · Microsoft Phi Safety Panels · MT-JailBench CrescendoX · OR-Bench · PacifAIst · PHARE · Qwen2 Safety Panel · RealityTest — Text AI-Identity Disclosure · RefusalBench · S-Eval · SafeDialBench · SALAD-Bench · Shell · SM-Bench · SORRY-Bench · SpeciEval · StereoTales Harmful Associations · SuperCLUE Safety · TAC · UAVBench safety-critical decision recognition · Vectara HHEM Factual Consistency · VETO Misfired Alignment · XSTest