Model profile
Voxtral Small 24B 2507
Evidence summary
Safety. Voxtral Small 24B 2507 does not meet the evidence gate for a Safety rank. Published Safety evidence spans 2 eval lineages and 2 of 7 components. Its strongest relative result is RealityTest — Text AI-Identity Disclosure (disclosure_probability, #5 of 17); its weakest is Cisco AI Defense Rolling Single-Turn Leaderboard (single_turn_attack_success_rate, #88 of 104).
Freedom. Voxtral Small 24B 2507 has an estimated Freedom rank of #166; its 90% source-sensitivity interval is #41–#638. Published Freedom evidence spans 1 eval lineages and 1 of 1 components. Its strongest relative result is Cisco AI Defense Rolling Single-Turn Leaderboard (single_turn_attack_success_rate, #17 of 104); its weakest is Cisco AI Defense Rolling Single-Turn Leaderboard (single_turn_attack_success_rate, #17 of 104).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Hugging Face ↗mistralai/Voxtral-Small-24B-2507
- Release source ↗repository created at proxy
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 |
|---|---|---|---|---|
| Cisco AI Defense Rolling Single-Turn Leaderboardsingle_turn_attack_success_rate | #88 / 104 | ↓69.94 | Source ↗official | |
| RealityTest — Text AI-Identity Disclosuredisclosure_probability | #5 / 17 | ↑0.541 | 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 |
|---|---|---|---|---|
| Cisco AI Defense Rolling Single-Turn Leaderboardsingle_turn_attack_success_rate | #17 / 104 | ↑69.94 | Source ↗official |