Command R+
Command R+.
Cohere Released 30 Aug 2024 26
Benchmarks with a score
0
Ranked first, out of five or more models
0
In the top three
Cohere Aug 2024 Command R+
Open weights
119.0
#200 of 271
Epoch Capabilities Index
- KOR-Bench 32.7% #26 /33
- Vectara Hallucination Leaderboard (HHEM) 5.40% #52 /148
- LMCA 5.01% #137 /139
Where it ranks
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
Knowledge and language
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Vectara Hallucination Leaderboard (2025 dataset) | 6.90% | #30 of 104 | Official leaderboard | |
| Vectara Hallucination Leaderboard (HHEM) | 5.40% | #52 of 148 | Official leaderboard | |
| MMLU | 69.4% | #58 of 127 | Epoch AI | |
| WildBench | 41.6 | #24 of 47 | Official leaderboard | |
| MixEval | 81.5% | #18 of 34 | Official leaderboard | |
| LMArena (Elo) | 1,276 | #247 of 333 | Official leaderboard | |
| ThaiExam | 37.9% | #18 of 24 | Official leaderboard | |
| RewardBench | 70.6% | #41 of 48 | Official leaderboard | |
| LongBench v2 | 31.6% | #31 of 33 | Official leaderboard |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| KOR-Bench | 32.7% | #26 of 33 | Official leaderboard | |
| DTBench | 24.9% | #150 of 178 | Official leaderboard | |
| SimpleBench | 17.4% | #98 of 100 | Official leaderboard |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| BigCodeBench-Hard | 13.2% | #78 of 96 | Official leaderboard |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Berkeley Function Calling Leaderboard (BFCL) | 48% | #53 of 67 | Official leaderboard | |
| ToolComp | 30.2% | #30 of 32 | Official leaderboard |
Professional work
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| LegalBench | 68.3% | #121 of 138 | Official leaderboard |
Safety and security
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| SimpleSafetyTests | 100% | #5 of 75 | Official leaderboard | |
| AgentDojo | 4.45% | #5 of 20 | Official leaderboard | |
| Anthropic Red Team | 98% | #56 of 75 | Official leaderboard | |
| HELM Safety | 86.0 | #62 of 75 | Official leaderboard | |
| AIR-Bench 2024 | 29.3% | #81 of 82 | Official leaderboard |
Other
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| HELM Lite | 44.0 | #45 of 71 | Official leaderboard | |
| LiveBench | 31.8% | #42 of 48 | Epoch AI | |
| Conceptual Reasoning Index | 11.6 | #145 of 149 | Official leaderboard | |
| ACCoRD | 18.0 | #146 of 150 | Official leaderboard | |
| LMCA | 5.01% | #137 of 139 | Epoch AI |
Cohere's models over time
Cohere's models in color, against every other model on the Epoch Capabilities Index, by release date.