When paired with Cohere Embed and Rerank, Command offers leading accuracy for advanced AI applications requiring information from documents and enterprise data sources.

Global businesses serve diverse people, organizations, and markets across a variety of languages. With Command, you can build applications that speak to the business world, whether it’s Arabic, Chinese (Mandarin), English, French, German, Italian, Japanese, Korean, Portuguese, or Spanish.

Connect Command to our Tool Use API to seamlessly automate business workflows. This capability helps upgrade AI applications from simple chatbots to powerful agents and research tools for increased productivity.

Command’s model outputs come with clear citations that mitigate the risk of hallucinations, and enable surfacing additional context from the source materials.

Command can write product descriptions, help draft emails, suggest example press releases, and much more.

Collaborate with a helpful co-author

Ask Command multiple questions about a document in order to assign a category to the document, extract a piece of information, or answer a general question about the document.

In-depth analysis of documents

Where answering a few questions about a document can save you a few minutes, doing it for thousands of documents can save a company years.

Millions of answers, at your command

SaaS

supporting

time to get started

benefits

limitations

pricing structure

Best for


Get started with LLMs instantly through Cohere's Software as a Service (SaaS) platform, all without managing backend infrastructure.


Private Deployments

Compute and infrastructure management required

Organizations with strict data residency requirements

Need for maximum control and customization

Organizations dealing with sensitive or customer data

Bring Cohere's models into your own infrastructure, whether in the cloud or on-premises. Maintain a high level of security and control over sensitive information.

For ultimate customization and control over security

Cloud AI Platforms

Deployment depends on cloud provider features

Companies seeking easy integration with cloud service

Use Cohere's models through a variety of platforms such as Amazon SageMaker, Amazon Bedrock, OCI Generative AI, etc.



For running Cohere’s models across cloud AI platforms

Cohere Command R is a family of highly scalable language models that balances high performance with strong accuracy. Together with Cohere Embed and Rerank models, Command enables companies to move beyond proof of concept and into production.

[block:html]
{
  "html": "<style>\n\n/*---Style the table to have no wrap---*/\n  \n  td:nth-child(1)  {\n\twhite-space: nowrap;\n  }\n</style> "
}
[/block]


Command R is a large language model optimized for conversational interaction and long context tasks. It targets the “scalable” category of models that balance high performance with strong accuracy, enabling companies to move beyond proof of concept and into production. 

Command R boasts high precision on [retrieval augmented generation](https://docs.cohere.com/docs/retrieval-augmented-generation-rag) (RAG) and tool use tasks, low latency and high throughput, a long 128,000-token context length, and strong capabilities across 10 key languages.

### Model Details

| Latest Model | Description                                                                                                                                                                                                                                                                                        | Max Tokens Context Length) | Endpoints                                     |
| ------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------- | --------------------------------------------- |
| `command-r`  | Command R is an instruction-following conversational model that performs language tasks at a higher quality, more reliably, and with a longer context than previous models. It can be used for complex workflows like code generation, retrieval augmented generation (RAG), tool use, and agents. | 128k                       | [Chat](https://docs.cohere.ai/reference/chat) |

## Unique Command R Model Capabilities

Command R has been trained on a massive corpus of diverse texts in multiple languages, and can perform a wide array of text-generation tasks. Moreover, Command R has been trained with a particular focus on excelling in some of the most critical business use-cases.

### Multilingual Capabilities

We want Command R to serve as many people, organizations, and markets as possible, so the new Command R is capable of interacting in many languages to a fairly high degree of accuracy.

The model is optimized to perform well in the following languages: English, French, Spanish, Italian, German, Brazilian Portuguese, Japanese, Korean, Simplified Chinese, and Arabic. 

Additionally, pre-training data has been included for the following 13 languages: Russian, Polish, Turkish, Vietnamese, Dutch, Czech, Indonesian, Ukrainian, Romanian, Greek, Hindi, Hebrew, Persian.

The model has been trained to respond in the language of the user. Here's an example:

```python
co.chat(
  message="Écris une description de produit pour une voiture électrique en 50 à 75 mots"
)
```

And here's what the response might look like:

```text
Découvrez la voiture électrique qui va révolutionner votre façon de conduire.
Avec son design élégant, cette voiture offre une expérience de conduite unique
avec une accélération puissante et une autonomie impressionnante. Sa
technologie avancée vous garantit une charge rapide et une fiabilité inégalée.
Avec sa conception innovante et durable, cette voiture est parfaite pour les 
trajets urbains et les longues distances. Profitez d'une conduite silencieuse
et vivez l'expérience de la voiture électrique!
```

Command R can not only be used to generate text in several languages but can also perform cross-lingual tasks such as translation or answering questions about content in other languages.

### Retrieval Augmented Generation

Command R has been trained with the ability to ground its generations. This means that it can generate responses based on a list of supplied document snippets, and it will include citations in its response indicating the source of the information.

For more information, check out our dedicated guide on [retrieval augmented generation](https://cohere-enterprise.readme.io/docs/retrieval-augmented-generation-rag).

### Tool Use

Command R has been trained with conversational tool use capabilities. Its tool use functionality takes a conversation as input (with an optional user-system preamble), along with a list of available tools. The model will then generate a json-formatted list of actions to execute on a subset of those tools. For more information, check out our dedicated [tool use](https://docs.cohere.com/docs/tool-use) guide.

<div class="rdmd-html"><style>

/*---Style the table to have no wrap---*/
  
  td:nth-child(1)  {
	white-space: nowrap;
  }
</style> </div>
<p>Command R is a large language model optimized for conversational interaction and long context tasks. It targets the “scalable” category of models that balance high performance with strong accuracy, enabling companies to move beyond proof of concept and into production. </p>
<p>Command R boasts high precision on <a href="https://docs.cohere.com/docs/retrieval-augmented-generation-rag" target="" title="">retrieval augmented generation</a> (RAG) and tool use tasks, low latency and high throughput, a long 128,000-token context length, and strong capabilities across 10 key languages.</p>
<h3 class="heading heading-3 header-scroll" align=""><div class="heading-anchor anchor waypoint" id="model-details"></div><div class="heading-text"><div id="section-model-details" class="heading-anchor_backwardsCompatibility"></div>Model Details</div><a aria-label="Skip link to Model Details" class="heading-anchor-icon fa fa-anchor" href="#model-details"></a></h3>

















<div class="rdmd-table"><div class="rdmd-table-inner"><table><thead><tr><th>Latest Model</th><th>Description</th><th>Max Tokens Context Length)</th><th>Endpoints</th></tr></thead><tbody><tr><td><button aria-label="Copy Code" class="rdmd-code-copy fa"></button><code class="rdmd-code lang- theme-light" data-lang="" name="" tabindex="0">command-r</code></td><td>Command R is an instruction-following conversational model that performs language tasks at a higher quality, more reliably, and with a longer context than previous models. It can be used for complex workflows like code generation, retrieval augmented generation (RAG), tool use, and agents.</td><td>128k</td><td><a href="https://docs.cohere.ai/reference/chat" target="" title="">Chat</a></td></tr></tbody></table></div></div>
<h2 class="heading heading-2 header-scroll" align=""><div class="heading-anchor anchor waypoint" id="unique-command-r-model-capabilities"></div><div class="heading-text"><div id="section-unique-command-r-model-capabilities" class="heading-anchor_backwardsCompatibility"></div>Unique Command R Model Capabilities</div><a aria-label="Skip link to Unique Command R Model Capabilities" class="heading-anchor-icon fa fa-anchor" href="#unique-command-r-model-capabilities"></a></h2>
<p>Command R has been trained on a massive corpus of diverse texts in multiple languages, and can perform a wide array of text-generation tasks. Moreover, Command R has been trained with a particular focus on excelling in some of the most critical business use-cases.</p>
<h3 class="heading heading-3 header-scroll" align=""><div class="heading-anchor anchor waypoint" id="multilingual-capabilities"></div><div class="heading-text"><div id="section-multilingual-capabilities" class="heading-anchor_backwardsCompatibility"></div>Multilingual Capabilities</div><a aria-label="Skip link to Multilingual Capabilities" class="heading-anchor-icon fa fa-anchor" href="#multilingual-capabilities"></a></h3>
<p>We want Command R to serve as many people, organizations, and markets as possible, so the new Command R is capable of interacting in many languages to a fairly high degree of accuracy.</p>
<p>The model is optimized to perform well in the following languages: English, French, Spanish, Italian, German, Brazilian Portuguese, Japanese, Korean, Simplified Chinese, and Arabic. </p>
<p>Additionally, pre-training data has been included for the following 13 languages: Russian, Polish, Turkish, Vietnamese, Dutch, Czech, Indonesian, Ukrainian, Romanian, Greek, Hindi, Hebrew, Persian.</p>
<p>The model has been trained to respond in the language of the user. Here&#x27;s an example:</p>
<div class="CodeTabs CodeTabs_initial theme-light"><div class="CodeTabs-toolbar" role="tablist"><button aria-selected="true" class="CodeTabs_active" role="tab" type="button">Python</button></div><div class="CodeTabs-inner" role="tabpanel"><pre><button aria-label="Copy Code" class="rdmd-code-copy fa"></button><code class="rdmd-code lang-python theme-light" data-lang="python" name="" tabindex="0">co.chat(
  message=&quot;Écris une description de produit pour une voiture électrique en 50 à 75 mots&quot;
)
</code></pre></div></div>
<p>And here&#x27;s what the response might look like:</p>
<div class="CodeTabs CodeTabs_initial theme-light"><div class="CodeTabs-toolbar" role="tablist"><button aria-selected="true" class="CodeTabs_active" role="tab" type="button">Text</button></div><div class="CodeTabs-inner" role="tabpanel"><pre><button aria-label="Copy Code" class="rdmd-code-copy fa"></button><code class="rdmd-code lang-text theme-light" data-lang="text" name="" tabindex="0">Découvrez la voiture électrique qui va révolutionner votre façon de conduire.
Avec son design élégant, cette voiture offre une expérience de conduite unique
avec une accélération puissante et une autonomie impressionnante. Sa
technologie avancée vous garantit une charge rapide et une fiabilité inégalée.
Avec sa conception innovante et durable, cette voiture est parfaite pour les 
trajets urbains et les longues distances. Profitez d&#x27;une conduite silencieuse
et vivez l&#x27;expérience de la voiture électrique!
</code></pre></div></div>
<p>Command R can not only be used to generate text in several languages but can also perform cross-lingual tasks such as translation or answering questions about content in other languages.</p>
<h3 class="heading heading-3 header-scroll" align=""><div class="heading-anchor anchor waypoint" id="retrieval-augmented-generation"></div><div class="heading-text"><div id="section-retrieval-augmented-generation" class="heading-anchor_backwardsCompatibility"></div>Retrieval Augmented Generation</div><a aria-label="Skip link to Retrieval Augmented Generation" class="heading-anchor-icon fa fa-anchor" href="#retrieval-augmented-generation"></a></h3>
<p>Command R has been trained with the ability to ground its generations. This means that it can generate responses based on a list of supplied document snippets, and it will include citations in its response indicating the source of the information.</p>
<p>For more information, check out our dedicated guide on <a href="https://cohere-enterprise.readme.io/docs/retrieval-augmented-generation-rag" target="" title="">retrieval augmented generation</a>.</p>
<h3 class="heading heading-3 header-scroll" align=""><div class="heading-anchor anchor waypoint" id="tool-use"></div><div class="heading-text"><div id="section-tool-use" class="heading-anchor_backwardsCompatibility"></div>Tool Use</div><a aria-label="Skip link to Tool Use" class="heading-anchor-icon fa fa-anchor" href="#tool-use"></a></h3>
<p>Command R has been trained with conversational tool use capabilities. Its tool use functionality takes a conversation as input (with an optional user-system preamble), along with a list of available tools. The model will then generate a json-formatted list of actions to execute on a subset of those tools. For more information, check out our dedicated <a href="https://docs.cohere.com/docs/tool-use" target="" title="">tool use</a> guide.</p>

Command R

We charge differently for input and output tokens. You are charged based on the sum of tokens processed. Fine-tuning inference costs will be the same as our base models.

Each text (input) classified counts as one classification. You are not charged for examples provided.

Embeddings perform best when the text to be embedded is less than 512 tokens. You can create up to 96 embeddings per API call.

We charge differently for input and output tokens. You are charged based on the sum of tokens processed. Only Command Light is available for fine-tuning at this moment. Fine-tuning inference costs will be the same as our base models.

We count a single search unit as a query with up to 100 documents to be ranked. Documents longer than 510 tokens when including the length of the search query will be split up into multiple chunks, where each chunk counts as a singular document. Fine-tuning inference costs will be the same as our base models.

We charge differently for input and output tokens. You are charged based on the sum of tokens processed.

Retrieval-augmented generation at production scale

Retrieval-augmented generation at production scale

Retrieval-augmented generation at production scale

Highly accurate for businesses

RAG-optimized models

Command R family of models

What’s possible with Command