Cohere — pricing & performance
Models served
11
of 1,299 tracked models
Mean output speed
97
tokens per second, averaged over
11 measured endpoints
11 measured endpoints
Mean latency
1.0s
time to first token, averaged over
11 measured endpoints
11 measured endpoints
Cheapest blended
$0.000
per 1M tokens at a 3:1
input:output blend
input:output blend
Output speed on Cohere
Output tokens per second for all 11 models on Cohere with published throughput telemetry, the most recent figure OpenRouter publishes for each host. Higher is better. We do not run these measurements ourselves.
The Known Good
11 models on Cohere
Full leaderboard →
| Model | Creator | Input $/M | Output $/M | Blended | Speed | TTFT | Context |
|---|---|---|---|---|---|---|---|
Cohere: Command A
|
Cohere | $2.50 | $10.00 | $4.38 | 18 | 2.0s | 256k |
|
Cohere: Command A+
|
Cohere | $0.30 | $1.50 | $0.60 | 88 | 0.5s | 192k |
|
Cohere: Command R (08-2024)
|
Cohere | $0.15 | $0.60 | $0.26 | 19 | 0.3s | 128k |
|
Cohere: Command R+ (08-2024)
|
Cohere | $2.50 | $10.00 | $4.38 | 24 | 0.3s | 128k |
|
Cohere: Command R7B (12-2024)
|
Cohere | $0.037 | $0.15 | $0.066 | 67 | 0.2s | 128k |
|
Cohere: North Mini Code (free)
|
Cohere | $0.000 | $0.000 | $0.000 | 62 | 0.6s | 256k |
DeepSeek: DeepSeek V4 Flash 0731
|
DeepSeek | $0.14 | $0.28 | $0.17 | 141 | 0.8s | 1.05M |
|
DeepSeek: DeepSeek V4.1 Flash
|
DeepSeek | $0.30 | $1.20 | $0.53 | 219 | 0.3s | 1.05M |
Tencent: Hy4 preview
|
Tencent | $0.83 | $2.50 | $1.25 | 36 | 4.7s | 262k |
| Mi Xiaomi: MiMo-V2.6-Flash | Xiaomi | $0.14 | $0.28 | $0.17 | 247 | 0.5s | 1.05M |
Z.ai: GLM 5.3 Flash
|
Z.ai | $0.15 | $0.50 | $0.24 | 145 | 0.6s | 1.05M |
The Known Good
Source. Prices, endpoint availability and throughput for Cohere are ingested from OpenRouter;
model metadata is enriched from models.dev. A blank cell means the figure was not published for that endpoint —
it never means zero. Blended price is (3 × input + output) ÷ 4, the same 3:1 blend used everywhere on this
site; see methodology. Quality scores are not provider-specific:
a model scores the same wherever it is hosted, so quality lives on the
model pages.