Models·Cohere

Command R (08-2024): Price, Context and Uses

cohere/command-r-08-2024

Cohere’s text model for multilingual retrieval, tool use, math, code and reasoning.

$0.15Input, per million tokens
$0.60Output, per million tokens
128KContext window
4KLongest answer

128K tokens of context — about 96,000 words, on a scale that ends at 2M.

What it can do

  • Tool use
  • Takes text
  • Gives text

Released 30 Aug 2024 · checked 3 Oct 2026.

About it

Command R (08-2024) is a Cohere text model released on August 30, 2024. It supports tool use and targets multilingual retrieval-augmented generation, where retrieved information helps shape an answer. The model also improves math, code and reasoning. It accepts text, returns text, and has a 128,000-token context window.

Review

In a line

Command R (08-2024) is Cohere’s text-only language model, created August 30, 2024. It updates the Command R model with improved multilingual retrieval and tool use.

What it's good at

This model suits multilingual retrieval-augmented generation (RAG), a setup that adds retrieved information to a response. It also targets math, code and reasoning tasks.

Tool use is supported. For more options, see models with tool calling.

Price in plain terms

Input costs $0.15 per million tokens. Output costs $0.6 per million tokens.

Reading a 500-word page costs under 0.01 cents. A chat turn with about 1,000 input tokens and 500 output tokens costs about 0.045 cents. At 1,000 such turns, the rough total is about $0.45. Compare it with cheapest language models.

Context and limits

The 128,000-token context window is how much text the model can take into account in one request. It can produce up to 4,000 tokens in one answer.

Inputs and outputs are text only, so it is not an image model.

Ways to use it

Use it where Command R (08-2024) is available for multilingual RAG and tool-based tasks.

The model identifier is cohere/command-r-08-2024. Other Cohere options are listed in Cohere models.

Verdict

Choose Command R (08-2024) when a text workflow needs a large working context, multilingual retrieval, tool use, or stronger math and code handling. Its token rates make short reading and chat jobs inexpensive at the stated examples.

Look elsewhere if you need images, non-text inputs or outputs, or answers longer than 4,000 tokens. Review the available Cohere models if this model’s text-only limits do not fit your application.

Verdict

Command R (08-2024) fits people building text workflows around retrieved information, especially across multiple languages. Its tool support and stated gains in math, code and reasoning broaden the jobs it can handle. The 128,000-token window gives room for long source material, while the example costs stay small for ordinary chat volumes. It is a poor match for image work or applications that need very long single responses, because it handles text only and stops at 4,000 output tokens.

  • 128,000-token context window
  • Supports tool use
  • Designed for multilingual retrieval-augmented generation
  • Improved math, code and reasoning
  • 1,000 sample chat turns cost about $0.45
  • Text inputs and outputs only
  • Single answers are limited to 4,000 tokens
  • Output tokens cost more than input tokens
  • Released August 30, 2024
  • Pick it ifMultilingual retrieval-augmented generation; Text apps that need tool use; Math, code and reasoning tasks
  • Skip it ifYou need image input or image output; Your answers regularly exceed 4,000 tokens

Pricing

The model charges separately for text sent in and text generated. Input is $0.15 per million tokens; output is $0.6 per million tokens. The supplied examples put a 500-word page under 0.01 cents and 1,000 typical chat turns at about $0.45.

Questions

What is Command R (08-2024)?

It is Cohere’s text-only language model, released August 30, 2024, for multilingual retrieval-augmented generation, tool use, math, code and reasoning.

How much does Command R (08-2024) cost?

Input costs $0.15 per million tokens and output costs $0.6 per million tokens.

How large is its context window?

It can take up to 128,000 tokens into account in one request.

What is the maximum answer length?

The maximum output is 4,000 tokens.

Can it process images?

No. Its inputs and outputs are text only.

What would 1,000 chat turns cost?

Using about 1,000 input tokens and 500 output tokens per turn, 1,000 turns cost about $0.45.

Does it support tools?

Yes. Tool use is listed as supported.

On these boards

More from Cohere