Command R: benchmark results
Cohere Command R model optimized for retrieval-augmented generation and enterprise chat. Provider: Cohere. Released 2024-03-11. Access: Open.
Aggregate Rating 1401 ± 14, rank #1223 of 1656 rated models, from 532 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MERA - SimpleAr | 99.4 | EM (%) | 73.8 |
| MERA - ruDetox | 30.14 | Joint Score (%) | 73.2 |
| HELM AIR-Bench 2024 - #6.22: Legal | 3.33 | Refusal Rate (%) | 72.7 |
| MERA - PARus | 89.2 | Accuracy (%) | 70.9 |
| MERA - MultiQ | 50.58 | F1 (%) | 68.7 |
| HELM AIR-Bench 2024 - #39-40.51: Classification of individuals - Social behaviors | 95 | Refusal Rate (%) | 67.4 |
| MERA - CheGeKa | 29.75 | F1 (%) | 67.3 |
| MERA - ruHateSpeech | 78.11 | Accuracy (%) | 60 |
| HELM AIR-Bench 2024 - #4.3: Adversely affects legal rights | 60 | Refusal Rate (%) | 58.1 |
| HELM AIR-Bench 2024 - #41-42.16: Unauthorized disclosure - Financial records | 100 | Refusal Rate (%) | 55.2 |
| MERA - ruTiE | 74.96 | Accuracy (%) | 55.2 |
| HELM NaturalQuestions (Open) | 72.05 | F1 (%) | 53.3 |
Interactive version: theaggregate.ai/model?slug=command-r · How It Works · Data refreshed daily, snapshot 2026-10-20.