llm-benchmarks
githubdrose.io

cohere Provider Benchmarks


Measured over the last 30 days

Models Tracked
8
Avg Tokens / Second
30.18
Avg Time to First Token (ms)
795.00
Last Updated
Sep 1, 2026

All cohere models

8 tracked · fastest first
cohere site
ModelAvg Toks/SecMinMaxAvg TTF (ms)
Command R7B (12-2024)53.2039.8061.60560.00
Command R7B (12-2024)43.7021.9065.00830.00
Command A38.0033.7049.50740.00
Command R (08-2024)33.6026.9038.90580.00
Command A31.102.7353.501090.00
Command R (08-2024)22.903.8049.80770.00
Command R+ (08-2024)10.201.8224.80900.00
Command R+ (08-2024)8.721.9119.30890.00

Frequently Asked Questions

Which cohere model is fastest?
Based on recent tests, Command R7B (12-2024) shows the highest average throughput among tracked cohere models.
How many recent measurements feed this dashboard?
This provider summary aggregates 168 individual prompts measured across 151 monitoring runs over the past month.