The GitHub Copilot model deprecation took effect on November 16, 2026 for three models: Claude Opus 4.8, Claude Opus 5 and GPT-5.6 Terra. In its changelog, GitHub said it removed these models from most Copilot features, including Copilot Chat, inline edits, ask and agent modes, and code completion. If you pinned one of these models in the model picker, you need to switch to a replacement.
GitHub recommends Claude Opus 5.5 to replace Claude Opus 4.8 and Claude Opus 5, and GPT-6.1 Sol or GPT-6 Luna to replace GPT-5.6 Terra.
In Business and Enterprise organizations, the replacement models turn on automatically unless an admin disabled the global default. No manual model removal is needed.
Individual annual plan subscribers can keep using Claude Opus 4.8. Claude Opus 5 was removed from the model picker only.
GitHub Copilot model deprecation: retired models and replacements
The table below shows the replacements GitHub listed. All three models were retired on November 16, 2026. According to GitHub, the deprecation was first announced on October 15.
| Retired model | Recommended replacement |
|---|---|
| Claude Opus 4.8 | Claude Opus 5.5 |
| Claude Opus 5 | Claude Opus 5.5 |
| GPT-5.6 Terra | GPT-6.1 Sol or GPT-6 Luna |

Claude models are made by Anthropic and GPT models by OpenAI. Copilot is the front door that lets you pick between them. GitHub's recommended replacements are the next models in the same family. We also cover how to use Claude Opus 5.5 in a coding tool in how to use Claude Opus 5.5 in Antigravity.
What changes on your plan
Individuals and organizations are affected differently. Here is the guidance from the GitHub changelog, split by plan.
| User | What changes | What to do |
|---|---|---|
| Individual plans (except annual plan subscribers) | All three models are retired | Switch to a replacement in the picker |
| Individual annual plan subscribers | Claude Opus 4.8 stays available | You can keep it, but compare it with Opus 5.5 |
| Business and Enterprise | Replacement models turn on automatically (unless an admin disabled the global default or blocked the model) | Admins check that the replacements are enabled in the model policy |
Claude Opus 5 left the model picker, but it remains as a runtime refusal fallback for access that already exists. That means you cannot select it yourself. Your organization's existing model policies still apply.
Organizations with the global default turned off do not get the replacements automatically. In that case, an admin must enable Claude Opus 5.5, GPT-6.1 Sol or GPT-6 Luna in the Copilot model policy before they show up in team members' model pickers.
Pricing of the replacement models
The replacement models differ a lot in price. According to the Copilot models and pricing docs, usage is measured in AI credits (1 credit = $0.01). Once you pass your plan limit, you are billed at the token rates below. A token is the unit a model uses to process text. Rates are per 1 million tokens.
| Model | Input | Output |
|---|---|---|
| Claude Opus 5.5 | $4 | $20 |
| GPT-6.1 Sol (input up to 272,000 tokens) | $2 | $10 |
| GPT-6 Luna (input up to 272,000 tokens) | $0.10 | $0.50 |
| Claude Opus 4.8 (for reference) | $5 | $25 |
For long jobs with more than 272,000 input tokens, GPT-6.1 Sol costs $4 for input and $15 for output, and GPT-6 Luna costs $0.20 for input and $0.75 for output. The retired Claude Opus 5 and GPT-5.6 Terra have no rates in this document. Going by the table alone, Opus 4.8 at $5 input and $25 output becomes $4 and $20 on Opus 5.5. Terra is not in the pricing docs, so you cannot compare its rate with Sol or Luna. After you switch, check the real cost on your usage screen.
According to the supported models docs, all three replacement models are generally available (GA) and are included in automatic model selection (Auto) in Copilot Chat, the CLI, the cloud agent and the Copilot app. They also support a 1 million token context and reasoning effort settings.
The fastest way to see which model fits your work is to compare them yourself. Send one everyday request (for example, adding tests to a single function) to two replacement models. Then compare the answer quality and the AI credits used side by side on the usage screen. A rate table alone cannot tell you the real cost, because each request uses a different number of tokens.
Steps to change your Copilot model
These steps are for VS Code. They follow the GitHub docs.
- Click the chat icon in the title bar to open Copilot Chat.
- Open the model dropdown at the bottom of the chat window.
- Pick one of Claude Opus 5.5, GPT-6.1 Sol or GPT-6 Luna.
- Send the same question once to confirm you get an answer.

If the replacement model does not appear in the list, check your organization policy first. According to the GitHub docs, Business plan members can change models only when their organization allows it. If the global default is off, an admin must enable the replacement models in the model policy. Settings for using external models are under Manage Models in the same doc. Here too, Business and Enterprise need the "Bring Your Own Language Model Key in Select IDEs" policy turned on.

If you use automation and integrations
GitHub told users to update any workflows and integrations that use a retired model name to a supported model. If your team writes model names in config files, search your repositories for the retired names (Opus 4.8, Opus 5, GPT-5.6 Terra) and make sure none are left. We cover how to use Copilot for automated tasks in GitHub Agentic Workflows is now generally available, and organization cost settings in GitHub Copilot code review pricing and organization billing settings. GitHub Enterprise customers can send questions to their account manager.
Frequently asked questions
Does my bill change when a Copilot model is retired?
The retirement itself does not change your bill. But the replacement models have different token rates, so after you switch, watch how your AI credits drop on the usage screen.
What happens if I do not change the model?
GitHub said no model removal work is needed, and the replacement models appear in the picker once they are enabled. If the model you picked was retired, just choose a new one in the picker.
Found this useful? Share it
Latest posts
Kakao Share Thumbnail Not Showing? Here's the Fix (JavaScript SDK 2025)
Learn why Kakao Share thumbnails don't appear even with a valid imageUrl, and fix it using scrapImage to convert your URL into a KAGE CDN URL. Step-by-step guide.
How to Generate CRM User Manuals in 20 Minutes with Claude Code Sub-Agents
Learn how Claude Code sub-agents can analyze your CRM codebase and auto-generate user manuals and visual guidebooks in just 20 minutes. See the real-world results.
GEO Generative Engine Optimization: The Complete 2025 AI Marketing Strategy Guide
Learn what GEO (Generative Engine Optimization) is, how it differs from traditional SEO, and discover 5 proven AI marketing strategies to get cited by AI search engines in 2025.
Related projects
Get new posts by email
Insights on marketing, analytics, and dev, delivered to your inbox.

