GitHub mark, a black circle with a white cat silhouette, the logo of GitHub, maker of GitHub Copilot

Copilot Model Deprecation: Power-User Guide to the November and December Retirements

by

Last updated: November 15, 2026

GitHub Copilot is retiring nine models in two waves. Claude Opus 4.8, Claude Opus 5 and GPT-5.6 Terra go on November 16, 2026. Claude Fable 5, Claude Haiku 4.5, Gemini 3.8 Flash, Gemini 4 Argon, Grok 4.6 and Kimi K3 follow on December 14. Switch to the suggested alternatives before each date.

This guide is based on GitHub’s December deprecation notice (November 13, 2026), the earlier mid-November notice (October 15, 2026), the supported models reference, the models and pricing page and the enterprise model availability guide. Prices are per million tokens and were current on November 15, 2026.

Contents

Deprecation schedule

GitHub says each deprecation applies across all Copilot experiences, including Copilot Chat, inline edits, ask and agent modes, and code completions. The two notices list these models and suggested alternatives:

Model Deprecation date Suggested alternative
Claude Opus 4.8 November 16, 2026 Claude Opus 5.5
Claude Opus 5 November 16, 2026 Claude Opus 5.5
GPT-5.6 Terra November 16, 2026 GPT-6.1 Sol or GPT-6 Luna
Claude Fable 5 December 14, 2026 Claude Fable 5.2
Claude Haiku 4.5 December 14, 2026 Claude Haiku 5.5
Gemini 3.8 Flash December 14, 2026 Gemini 4 Hydrogen
Gemini 4 Argon December 14, 2026 Gemini 4 Carbon
Grok 4.6 December 14, 2026 Grok 4.7
Kimi K3 December 14, 2026 To be announced

This is the latest in a steady series. The retirement history on the supported models page shows earlier waves on September 1, October 2 and October 19, 2026, when models such as GPT-5.4, GPT-5.5 and Claude Opus 4.7 left Copilot.

How the replacements compare on price

Copilot now bills model use in GitHub AI Credits, where one credit equals USD 0.01, at per-token rates once included allowances run out. The pricing page lists these default-tier rates per million tokens:

Retiring model (input / cached / output) Suggested alternative (input / cached / output) Change
Claude Opus 4.8 or Opus 5: $5.00 / $0.50 / $25.00 Claude Opus 5.5: $4.00 / $0.20 / $20.00 Cheaper
GPT-5.6 Terra: $2.00 / $0.20 / $12.00 GPT-6.1 Sol: $2.00 / $0.10 / $10.00; GPT-6 Luna: $0.10 / $0.01 / $0.50 Same or cheaper
Claude Fable 5: $10.00 / $1.00 / $50.00 Claude Fable 5.2: $8.00 / $0.20 / $40.00 Cheaper
Claude Haiku 4.5: $1.00 / $0.10 / $5.00 Claude Haiku 5.5: $0.10 / $0.01 / $0.50 up to 100K input tokens Much cheaper
Gemini 3.8 Flash: $0.75 / $0.075 / $3.75 (promotional) Gemini 4 Hydrogen: $1.50 / $0.075 / $7.50 About twice the input and output price
Gemini 4 Argon: $2.00 / $0.10 / $10.00 (promotional) Gemini 4 Carbon: $2.00 / $0.10 / $10.00 until December 1, then $4.00 / $0.20 / $20.00 Doubles on December 1
Grok 4.6: $2.00 / $0.50 / $6.00 Grok 4.7: $2.00 / $0.50 / $6.00 Same
Kimi K3: $3.00 / $0.30 / $15.00 To be announced Not yet known

Most moves lower your per-token cost. The two Gemini moves are the exception: Gemini 4 Hydrogen costs twice as much per token as Gemini 3.8 Flash, and Gemini 4 Carbon’s promotional price ends on December 1, two weeks before Gemini 4 Argon retires. Cache write charges are not shown. Code completions and next edit suggestions are not billed in AI credits, so retiring a chat model does not change completion costs.

For background on the newer models, see our guides to Gemini 4 Hydrogen, Claude Fable 5.2 and Claude Haiku 5.5.

How to switch models as an individual

On individual plans you pick models yourself; there are no policies to manage. In Visual Studio Code, the model change guide describes this process:

  1. Open Copilot Chat from the chat icon in the title bar.
  2. Check the current model in the dropdown at the bottom of the chat view. If it names a retiring model, note which task you use it for.
  3. Select the suggested alternative from the same dropdown, or choose Copilot Fusion with an effort level to let Copilot pick the model.
  4. Update saved workflows. GitHub asks users to update workflows and integrations that name a retiring model before each date.
  5. Compare costs with the table above, especially if you rely on Gemini 3.8 Flash for cheap, high-volume chats.

Changing the chat model does not change the model used for inline suggestions. For the routing option, see our GitHub Copilot Fusion guide, and for terminal use, our Copilot CLI 1.0.111 guide.

Steps for Business and Enterprise admins

GitHub says that under default model enablement, the suggested alternatives are enabled automatically for Copilot Business and Enterprise unless an administrator turned off the global default or explicitly disabled the model. Admins who turned the default off need to enable the alternatives themselves. According to the enterprise guide, enterprise owners do this as follows:

  1. Go to the enterprise on GitHub.com and click AI controls at the top of the page.
  2. In the sidebar, click Copilot, then Configure models.
  3. Set each suggested alternative to Enabled, or delegate it to organizations or enterprise teams.
  4. Check models labeled Delegate to Default Policy, which follow the Default availability policy.
  5. If you use Copilot Fusion, open Configure Copilot Fusion models and enable the new models there too.

Organization owners make the same change under the organization’s Settings, then Copilot, then Models. GitHub says no action is needed to remove the deprecated models.

Exceptions and caveats

  • Annual plans. Claude Opus 4.8 and Claude Haiku 4.5 stay available to individual subscribers on annual plans.
  • Claude Fable 5.2 is not enabled by default. The supported models page says Fable models fall outside default enablement because they are not covered by GitHub’s data retention agreement. Admins must enable Fable 5.2 explicitly.
  • Kimi K3 has no named successor. It is also an open-weight model that is disabled by default for organizations. Teams that rely on it should test another model before December 14.
  • The reference pages lag. On November 15, the supported models page still listed the November models as generally available. Use the two changelog notices for dates.

For a similar migration on the Gemini API, see our Gemini 2.5 deprecation guide.

Frequently asked questions

Which GitHub Copilot models are being deprecated?

On November 16, 2026, GitHub deprecates Claude Opus 4.8, Claude Opus 5 and GPT-5.6 Terra. On December 14, 2026, it deprecates Claude Fable 5, Claude Haiku 4.5, Gemini 3.8 Flash, Gemini 4 Argon, Grok 4.6 and Kimi K3 across all Copilot experiences.

What should I use instead of Claude Haiku 4.5 in Copilot?

GitHub suggests Claude Haiku 5.5. On Copilot’s pricing page it costs USD 0.10 per million input tokens and USD 0.50 per million output tokens up to 100K input tokens, compared with USD 1.00 and USD 5.00 for Claude Haiku 4.5.

Is there a replacement for Kimi K3 in GitHub Copilot?

Not yet. GitHub lists the suggested alternative for Kimi K3 as “To be announced.” Kimi K3 is also an open-weight model that is disabled by default for Copilot Business and Enterprise, so organizations that use it should plan a fallback before December 14.

Do I keep Claude Haiku 4.5 on an annual Copilot plan?

Yes. GitHub says Claude Haiku 4.5 remains available to individual Copilot subscribers on annual plans, and the December deprecation does not apply to them. Claude Opus 4.8 likewise stays available to individual annual subscribers after November 16.

Do admins need to remove deprecated models?

No. GitHub says no action is required to remove models after deprecation. Admins should instead make sure the suggested alternatives are enabled, especially if the default availability policy is turned off or a model sits outside default enablement.

Sources

Related answers

Leave a Comment