Command R7B: benchmark results

Provider: Cohere. Released 2024-12-13. Access: Open.

Unified ELO 1415 ± 37, rank #1790 of 2656 rated models, from 31 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
WMT25 English-Maasai chrF++24.3chrF++84.6
BFCL V4 - Web Search27Accuracy (%)67.1
BFCL V4 - Irrelevance Detection81.65Accuracy (%)48.8
Gorilla API Bench (BFCL)32.07Overall Accuracy (%)46.3
BFCL V4 - Live69.06Accuracy (%)39
BFCL V4 - Multi Turn8.25Accuracy (%)34.1
WMT25 English-Maasai ESA1.6ESA Human Score29.4
BFCL V4 - Non-Live AST80.96Accuracy (%)28
BFCL V4 - Memory5.16Accuracy (%)19.5
BFCL V4 - Relevance Detection68.75Accuracy (%)19.5
WMT25 English-Bhojpuri chrF++22.9chrF++13.2
AI for Education Pedagogy - Technology60.38Accuracy (%)9.8

Interactive version: theaggregate.ai/model?slug=command-r7b · How It Works · Data refreshed daily, snapshot 2026-09-19.