Model profile
Command R7B
Evidence summary
Safety. Command R7B has an estimated Safety rank of #333; its 90% source-sensitivity interval is #297–#340. Its behavior-only rank is #341; company governance moves the combined estimate to #333. Published Safety evidence spans 3 eval lineages and 4 of 7 components. Its strongest relative result is Enkrypt AI Safety Leaderboard (toxicity_attack_non_success_rate, #161 of 268); its weakest is Enkrypt AI Safety Leaderboard (bias_attack_non_success_rate, #270 of 270).
Freedom. Command R7B has an estimated Freedom rank of #55; its 90% source-sensitivity interval is #28–#342. Published Freedom evidence spans 1 eval lineages and 1 of 1 components. Its strongest relative result is Enkrypt AI Safety Leaderboard (cbrn_attack_non_success_rate, #42 of 270); its weakest is Enkrypt AI Safety Leaderboard (toxicity_attack_non_success_rate, #108 of 268).
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 · cohere · 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 |
|---|---|---|---|---|
| AgentDrive Safety Compliancescr | #38 / 48 | ↑57.5 | Source ↗official | |
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #270 / 270 | ↑0.52 | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #229 / 270 | ↑71.33 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #223 / 270 | ↑45 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #161 / 268 | ↑95.55 | Source ↗official | |
| HUMAINE Trust, Ethics and Safetytrust_ethics_safety_score | #54 / 54 | ↑20.03 | 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 |
|---|---|---|---|---|
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #42 / 270 | ↓71.33 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #48 / 270 | ↓45 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #108 / 268 | ↓95.55 | Source ↗official |