Model profile
Devstral 2
Evidence summary
Safety. Devstral 2 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 AA-Omniscience (hallucination_rate, #213 of 345); its weakest is Cisco AI Defense Rolling Single-Turn Leaderboard (single_turn_attack_success_rate, #100 of 104).
Freedom. Devstral 2 has an estimated Freedom rank of #49; its 90% source-sensitivity interval is #3–#408. 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, #5 of 104); its weakest is Cisco AI Defense Rolling Single-Turn Leaderboard (single_turn_attack_success_rate, #5 of 104).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗devstral-2
- 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 |
|---|---|---|---|---|
| AA-Omnisciencehallucination_rate | #213 / 345 | ↓0.8527 | Source ↗official | |
| Cisco AI Defense Rolling Single-Turn Leaderboardsingle_turn_attack_success_rate | #100 / 104 | ↓79.76 | 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 | #5 / 104 | ↑79.76 | Source ↗official |