No available models for github copilot cli after an update

[neilyan-msft]

Select Topic Area

Question

Copilot Feature Area

General

Body

I have updated github copilot cli to the latest version 0.0.349. It shows no available model. The github copilot chat is working well. The issue only happens to copilot cli.

[wassupjay]

The “no available model” error in GitHub Copilot CLI 0.0.349 is a known issue caused by a backend mismatch—update your gh CLI and reauthenticate with gh auth refresh -h github.com -s copilot to fix it.

[neilyan-msft]

Thank you @wassupjay , I have tried that, it doesn't work. My gh is the latest version

gh --version
gh version 2.82.1 (2025-10-22)

I think the copilot is also the latest version:

copilot
 Welcome to GitHub Copilot CLI
 Version 0.0.349 · Commit 3469b3e

I refreshed the token by using your command. I still get the error 'no available model`

[neilyan-msft]

Anyone is looking at this issue? Just updated copilot to a newer version. 0.0.353. The model list appeared for a second, then disappeared.

[vishwaswami24]

Solution 1: Check Organization Policy (Most Likely Fix)
If you are using GitHub Copilot as part of a company or organization (i.e., you have a "Copilot Business" or "Copilot Enterprise" seat), you will need to contact your organization's administrator.

What to tell your administrator:

"The GitHub Copilot CLI policy is currently disabled for our organization. To fix this, please:

Navigate to the organization's settings: https://github.com/organizations/[YOUR_ORG_NAME]/settings/copilot/policies

Find the policy for "Copilot CLI" or "Use of GitHub Copilot in the CLI".

Enable this policy."

This is the most common solution for the "no available model" error when other Copilot services are working.

Solution 2: Try to Manually Select a Model (Quick Workaround)

There seems to be a bug in some CLI versions where the list of models doesn't load correctly on startup. You may be able to force it to refresh by attempting to set a model, even if you don't know the model names.

In your terminal, run the copilot command to start the interactive session.

Once it's running, type the following command:

/model gpt-4
Even if this command fails, some users have reported that just attempting to set a model (e.g., /model gpt-4, /model claude-3, etc.) can kick-start the connection and make the models appear.

Solution 3: Re-authenticate the CLI

It's possible the CLI's authentication token is stale or doesn't have the correct permissions, even if your IDE's token is fine.

Run the following command to log out:
copilot auth logout

Then, log back in. This will force a new authentication flow through your browser.
copilot auth login