ModelsCommand

Command and Command Light

Note

For most use cases we recommend our latest model Command R instead.

Model Details

Latest ModelDescriptionModalityContext LengthMaximum Output TokensEndpoints
commandAn instruction-following conversational model that performs language tasks with high quality, more reliably and with a longer context than our base generative models.Text4k4kChat,
Summarize
command-lightA smaller, faster version of command. Almost as capable, but a lot faster.Text4k4kChat,
Summarize
command-nightlyTo reduce the time between major releases, we put out nightly versions of command models. For command, that is command-nightly.

Be advised that command-nightly is the latest, most experimental, and (possibly) unstable version of its default counterpart. Nightly releases are updated regularly, without warning, and are not recommended for production use.		Text128K4kChat
command-light-nightlyTo reduce the time between major releases, we put out nightly versions of command models. For command-light, that is command-light-nightly.

Be advised that command-light-nightly is the latest, most experimental, and (possibly) unstable version of its default counterpart. Nightly releases are updated regularly, without warning, and are not recommended for production use.		Text4k4kChat

The Command family of models responds well with instruction-like prompts, and are available in two variants: command-light and command. The command model demonstrates better performance, while command-light is a great option for applications that require fast responses.

To reduce the turnaround time for releases, we have nightly versions of Command available. This means that every week, you can expect the performance of command-nightly and command-light-nightly to improve.

For information on toxicity, safety, and using this model responsibly check out our Command model card.

Example Prompts

Get Started

Set up

Install the SDK, if you haven’t already.

pip install cohere

Then, set up the Cohere client.

PYTHON
1import cohere  
2co = cohere.ClientV2(api_key="<YOUR API KEY>")

Create prompt

PYTHON
1message = "Write an introductory paragraph for a blog post about language models."

Generate text

PYTHON
1response = co.chat(model="command",
2                   messages=[{"role" : "user", "content" : message}]
3)
4
5intro_paragraph = response.message.content[0].text

FAQ

Can users train Command?

Users cannot train Command in OS at this time. However, our team can handle this on a case-by-case basis. Please email team@cohere.com if you’re interested in training this model.

Where can I leave feedback about Cohere generative models?

Please leave feedback on Discord.

What’s the context length on the command models?

A model’s “context length” refers to the number of tokens it’s capable of processing at one time. In the table above, you can find the context length (and a few other relevant parameters) for the different versions of the command models.