Enable Copilot Enterprise Model Policy for New AI Models

When GitHub deprecates a model, it does not automatically push the replacement into your Copilot Enterprise tenant. As of October 19, 2026, GitHub deprecated several models across most GitHub Copilot experiences, including Copilot Chat, inline edits, ask and agent modes, and code completions. To use the suggested alternatives, Copilot Enterprise administrators must enable access to those alternative models through their model policies in Copilot settings.

You can verify availability by checking your individual Copilot settings and confirming that the policy is enabled for the specific model. Once enabled, you will see the model in the Copilot Chat model selector in VS Code and on github.com. No action is required to remove the deprecated models from your tenant.

What changed in the October 19 deprecation#

GitHub announced these deprecations on September 18, 2026. They took effect on October 19, 2026. You must update your workflows and integrations to use the supported models.

Here is the full list of deprecated models and their suggested alternatives:

You can check the GitHub changelog for the latest updates.

Step by step: Enable access to replacement models#

Copilot Enterprise uses model policies to control which AI models your organization can use. This gatekeeping prevents unexpected billing spikes and ensures developers only access approved models. When GitHub rolls out a new model like GPT-6.1 Sol or GPT-6 Luna, the default policy keeps it hidden until an administrator explicitly enables it.

To enable access for your organization, follow these steps as an administrator: 1. Sign in to github.com with your Enterprise administrator account. 2. Navigate to your Enterprise account settings. 3. Locate the Copilot settings section. 4. Open the Model policies tab. 5. Find the newly available models, such as GPT-6.1 Sol or GPT-6 Luna. 6. Enable the policy for the specific model you want to roll out. 7. Save your changes.

If you manage multiple developers, you can use this simple Python script to generate a mapping of deprecated models to their replacements. This helps when auditing your internal documentation.

deprecated_models = {
    "Gemini 3.7 Flash": "Gemini 3.8 Flash",
    "GPT-5.5": "GPT-6.1 Sol",
    "GPT-5.4": "GPT-6.1 Sol",
    "GPT-5.4 mini": "GPT-6 Luna",
    "GPT-5 mini": "GPT-6 Luna",
    "Grok 4.5": "Grok 4.7",
}

def get_replacement(model_name):
    if model_name in deprecated_models:
        return f"Replace {model_name} with {deprecated_models[model_name]}"
    return f"{model_name} is not deprecated."

for model in ["GPT-5.5", "GPT-5.4 mini", "GPT-6 Astra"]:
    print(get_replacement(model))

Choosing your replacement model: Sol vs Luna#

If you previously used GPT-5.5 or GPT-5.4, GitHub suggests GPT-6.1 Sol as the alternative. If you used GPT-5.4 mini or GPT-5 mini, GitHub suggests GPT-6 Luna. For a deeper look, read our breakdown of GPT-6.1 Sol vs GPT-6 Luna Copilot.

According to OpenAI's models page, GPT-6.1 Sol is intended to balance intelligence and cost. It has a model ID of gpt-6.1-sol, a context window of 1.05M tokens, and a max output of 128K tokens. Its knowledge cutoff is April 30, 2026. It supports functions, web search, file search, and computer use. Reasoning levels include low, medium, high, xhigh, and max.

GPT-6 Luna is OpenAI's most efficient model for focused, high-volume tasks. It has a model ID of gpt-6-luna, a context window of 1.05M tokens, and a max output of 128K tokens. Its knowledge cutoff is May 18, 2026. It supports the same tools. Reasoning levels include none, low, medium, high, xhigh, and max.

What it costs to run the new models#

When you enable a new model policy, you should understand the pricing implications. OpenAI's pricing page lists the standard rates for these models. GPT-6.1 Sol costs $2.00 per million input tokens and $10.00 per million output tokens on the Standard tier. GPT-6 Luna costs $0.10 per million input tokens and $0.50 per million output tokens on the Standard tier.

If you run large batch jobs, OpenAI lists Batch pricing too. GPT-6.1 Sol Batch costs $1.00 per million input tokens and $5.00 per million output tokens. GPT-6 Luna Batch costs $0.05 per million input tokens and $0.25 per million output tokens.

For faster processing, OpenAI offers Fast and Ultrafast tiers. GPT-6.1 Sol Fast costs $4.00 per million input tokens and $20.00 per million output tokens. GPT-6.1 Sol Ultrafast costs $12.00 per million input tokens and $60.00 per million output tokens. You can also use GPT-6 Luna Fast for $0.20 per million input tokens and $1.00 per million output tokens.

Update your workflows and API integrations#

GitHub noted that you should update your workflows and integrations to use supported models. This means any script, GitHub Action, or internal tool that references the old model names needs an update. If your internal tools call Copilot APIs directly, calling a deprecated model ID will fail. You must swap the old IDs for the new ones.

For example, if your code uses the Responses API to chain turns, you might need to point it to the new model IDs. You can pass the model ID string directly in your API calls. See our guide on using the Responses API previous_response_id to chain turns in Python.

If you have questions or concerns, GitHub Enterprise customers are encouraged to reach out to their account manager for further assistance. You can also join the GitHub Community to share your feedback.

Verify availability in VS Code and github.com#

Once an administrator enables the policy, the new models become available in the Copilot Chat model selector in VS Code and on github.com. If you are a developer and you do not see the new models, confirm that your organization administrator enabled the policy. You can also check your individual Copilot settings to confirm the policy is enabled for the specific model.

To see the model selector in VS Code: 1. Open VS Code. 2. Open the Copilot Chat panel. 3. Look for the model dropdown menu at the top of the chat panel. 4. Select the newly enabled model, such as GPT-6.1 Sol or GPT-6 Luna.

If the model is still missing, try reloading the VS Code window. You can do this by opening the command palette and running the Developer: Reload Window command.

Common mistakes and gotchas#

1. Waiting for deprecated models to disappear. You do not need to take action to remove the deprecated models from your tenant.
2. Assuming the new models are enabled by default. They are not. You must enable access to alternative models through your model policies in Copilot settings.
3. Forgetting to verify the policy. As an administrator, you can verify availability by checking your individual Copilot settings and confirming that the policy is enabled for the specific model.
4. Not reloading VS Code. Once enabled, the model appears in the Copilot Chat model selector in VS Code and on github.com. If the model does not appear, try reloading VS Code.
5. Ignoring API integrations. If your internal tools call Copilot APIs directly, calling a deprecated model ID will fail. Update your code to use the new IDs.

If your company uses multiple AI platforms, keeping track of settings can be tough. You might also need to export Claude instructions for enterprise audits to stay compliant.

Quick reference checklist#

• Check the GitHub changelog for the latest model deprecations.
• Map your deprecated models to their suggested alternatives.
• Sign in to github.com as a Copilot Enterprise administrator.
• Go to Copilot settings and open Model policies.
• Enable the policy for GPT-6.1 Sol, GPT-6 Luna, or other replacements.
• Verify the model appears in the Copilot Chat model selector in VS Code and on github.com.
• Update internal API integrations to use the new model IDs.

Quick answers