Model profile
Command R
Coheredeveloper
2024-03-11release date
#207 / 305overall rank
8eval lineages
Evidence summary
Command R has an estimated overall rank of #207; its 90% source-sensitivity interval is #55–#293. Its behavior-only rank is #210; company governance moves the combined estimate to #207. Published evidence spans 8 evals and 4 of 7 behavior components. Its strongest relative result is AgentDojo (targeted_attack_success_rate, #1 of 15); its weakest is AIRBench 2024 Safety Scenarios (safety_scenarios, #78 of 80).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- OpenRouter ↗cohere/command-r
- Official model documentation ↗Family-level model document · cohere · first party
- Release source ↗direct research preview identity date
Published eval results
Rank is within that sub-eval. Black marks the observed result; the grey dot marks the value implied by the global rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Source |
|---|---|---|---|---|
| AgentDojotargeted_attack_success_rate | #1 / 15 | ↓0.0334 | Source ↗official | |
| AgentDojoutility_under_attack | #13 / 15 | ↑0.3084 | Source ↗official | |
| AIRBench 2024 Safety Scenariossafety_scenarios | #78 / 80 | ↑0.318 | Source ↗official | |
| Cisco AI Defense Rolling Single-Turn Leaderboardsingle_turn_attack_success_rate | #78 / 104 | ↓62.91 | Source ↗official | |
| HELM Safetyanthropic_red_team | #73 / 80 | ↑0.937 | Source ↗official | |
| HELM Safetybbq | #73 / 80 | ↑0.724 | Source ↗official | |
| HELM Safetyharmbench | #70 / 80 | ↑0.504 | Source ↗official | |
| HELM Safetysimple_safety_tests | #69 / 80 | ↑0.943 | Source ↗official | |
| HELM Safetyxstest | #57 / 80 | ↑0.939 | Source ↗official |