Providers / Cohere

Cohere AI models

Enterprise-focused lab known for the Command models plus embeddings and rerank.

Lifecycle watch · 10
The lineup

All Cohere models

Official pricing ↗
Model ↕Status ↕Context ↕Input $/M ↕Output $/M ↕Blended $/M ↕Cutoff ↕
Command R7Bopen
Cohere · text
GA128K$0.037$0.15$0.066—
Command R+ (08-2024)open
Cohere · text
GA128K$2.50$10$4.38—
Command R (08-2024)open
Cohere · text
GA128K$0.15$0.6$0.262—
Embed 4 (embed-v4.0)
Cohere · text, image
GA128K$0.12———
Command A Plusopen
Cohere · text, image
GA128K————
Command Aopen
Cohere · text
GA256K————
Command A Reasoningopen
Cohere · text
GA256K————
Command A Visionopen
Cohere · text, image
GA128K————
Command A Translateopen
Cohere · text
GA8K————
Embed English v3.0
Cohere · text, image
GA512————
Embed English Light v3.0
Cohere · text, image
GA512————
Embed Multilingual v3.0
Cohere · text, image
GA512————
Embed Multilingual Light v3.0
Cohere · text, image
GA512————
Rerank 4 Pro (rerank-v4.0-pro)
Cohere · text
GA32K————
Rerank 4 Fast (rerank-v4.0-fast)
Cohere · text
GA32K————
Rerank 3.5 (rerank-v3.5)
Cohere · text
GA4K————
Rerank English v3.0
Cohere · text
GA4K————
Rerank Multilingual v3.0
Cohere · text
GA4K————
Command R+ (04-2024)open
Cohere · text
Deprecated128K$3$15$6—
Command R (03-2024)open
Cohere · text
Deprecated128K$0.5$1.50$0.75—
Command (legacy)
Cohere · text
Deprecated4K$1$2$1.25—
Command Light (legacy)
Cohere · text
Deprecated4K$0.3$0.6$0.375—
Rerank English v2.0
Cohere · text
Retired—————
Rerank Multilingual v2.0
Cohere · text
Retired—————
Embed English v2.0
Cohere · text
Retired—————
Embed English Light v2.0
Cohere · text
Retired—————
Embed Multilingual v2.0
Cohere · text
Retired—————
Summarize (legacy endpoint)
Cohere · text
Retired—————

Blended = 0.75 × input + 0.25 × output $/M tokens (a fair single-number cost proxy). Click any header to sort.

FAQ

Cohere pricing & models

What is the cheapest Cohere model?

Command R7B is the cheapest generally-available Cohere model we track, at $0.037 per 1M input tokens and $0.15 per 1M output tokens ($0.066/1M blended).

What is Cohere's flagship model?

Command R7B is Cohere's most prominent model in our catalog, with a 128K-token context window and pricing of $0.037/$0.15 per 1M input/output tokens.

How many Cohere models are there?

We track 28 Cohere models, of which 18 are generally available and 10 are deprecated or scheduled for retirement.

Which Cohere models are being deprecated?

Rerank English v2.0 (retires 30 Apr 2025), Rerank Multilingual v2.0 (retires 30 Apr 2025), Summarize (legacy endpoint) (retires 15 Sep 2025), Embed English v2.0 (retires 4 Apr 2026), Embed English Light v2.0 (retires 4 Apr 2026), Embed Multilingual v2.0 (retires 4 Apr 2026), Command R+ (04-2024), Command R (03-2024), Command (legacy), Command Light (legacy).