Command R7B: benchmark results
Cohere's Command R7B model, part of the Command R series, released under a CC-BY-NC-4.0 license in December 2024; the linked model card provides no further architecture or training details. Provider: Cohere. Released 2024-12-13. Access: Open.
Aggregate Rating 1351 ± 22, rank #7667 of 10916 rated models, from 23 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| ContextEcho | -15 | Delta (self-reported) | 100 |
| WMT25 English-Maasai chrF++ | 24.3 | chrF++ | 84.6 |
| WMT25 English-Maasai ESA | 1.6 | ESA Human Score | 29.4 |
| WMT25 English-Bhojpuri chrF++ | 22.9 | chrF++ | 13.2 |
| AI for Education Pedagogy - Technology | 60.38 | Accuracy (%) | 9.8 |
| PLCC - Geography | 33 | Accuracy (%) | 7.8 |
| AI for Education SEND | 53.21 | Accuracy (%) | 6.3 |
| AI for Education Pedagogy - Primary | 47.89 | Accuracy (%) | 5.9 |
| AI for Education Pedagogy | 46.05 | Accuracy (%) | 5.7 |
| AI for Education Pedagogy - Secondary | 45.75 | Accuracy (%) | 5.7 |
| AI for Education Pedagogy - Maths | 40.48 | Accuracy (%) | 5.3 |
| AI for Education Pedagogy - Science | 40.98 | Accuracy (%) | 4.9 |
Interactive version: theaggregate.ai/model?slug=command-r7b · How It Works · Data refreshed daily, snapshot 2026-11-08.