Allow organizations to disable GPT-5.3-Codex fallback or cap fallback costs

[kosetei]

🏷️ Discussion Type

Product Feedback

💬 Feature/Topic Area

Copilot in GitHub

Body

Summary

GitHub Copilot uses GPT-5.3-Codex as a base fallback model, including for some tasks handled through automatic model selection. According to the documentation, this fallback cannot be disabled through enterprise or organization model policies.

This creates a serious cost-control problem when a user explicitly selects a model that is both preferred and less expensive, such as GPT-6 Luna.

Problem

In my case, I explicitly selected GPT-6 Luna, but approximately 1,500 AI credits were recorded as GPT-5.5 usage. I have already opened a support ticket regarding that specific billing issue.

More broadly, the documented fallback behavior raises the following concern:

• The user explicitly selects GPT-6 Luna.
• The organization may configure model policies.
• GitHub can internally substitute GPT-5.3-Codex as a fallback.
• The fallback cannot be disabled by the user or administrator.
• The substituted model may consume more AI credits than the selected model.

If customers cannot prevent or approve this substitution, they should not be charged at the higher fallback-model rate.

Requested changes

GitHub should implement at least one of the following:

1. Do not charge additional AI credits when an internal fallback model is used.
2. Never fall back to a model that costs more than the explicitly selected model.
3. Allow enterprise and organization administrators to disable fallback models.
4. Notify the user before using a higher-cost fallback model.
5. Clearly show the selected model and the actually executed model in usage records.
6. Allow administrators to set a maximum cost multiplier for fallback execution.

Why this matters

Enterprise customers require predictable costs and enforceable model policies.

A model-selection interface cannot function as a meaningful cost-control mechanism if GitHub can silently replace the selected model with a more expensive model that customers cannot disable.

Fallback behavior for service availability is understandable, but the additional cost of an unavoidable internal fallback should not be transferred to the customer without notice or consent.

Related discussion

• GitHub Copilot AI Credits Are Unfair, Expensive, and Killing Real Development Workflows #198015 discusses the broader issues of AI Credits pricing, predictability, and cost transparency.
• This request focuses specifically on unavoidable fallback to a higher-cost model that the user and organization cannot disable.

[github-actions[bot]]

💬 Your Product Feedback Has Been Submitted 🎉

Thank you for taking the time to share your insights with us! Your feedback is invaluable as we build a better GitHub experience for all our users.

Here's what you can expect moving forward ⏩

• Your input will be carefully reviewed and cataloged by members of our product teams. 
  • Due to the high volume of submissions, we may not always be able to provide individual responses.
  • Rest assured, your feedback will help chart our course for product improvements.
• Other users may engage with your post, sharing their own perspectives or experiences. 
• GitHub staff may reach out for further clarification or insight. 
  • We may 'Answer' your discussion if there is a current solution, workaround, or roadmap/changelog post related to the feedback.

Where to look to see what's shipping 👀

• Read the Changelog for real-time updates on the latest GitHub features, enhancements, and calls for feedback.
• Explore our Product Roadmap, which details upcoming major releases and initiatives.

What you can do in the meantime 💻

• Upvote and comment on other user feedback Discussions that resonate with you.
• Add more information at any point! Useful details include: use cases, relevant labels, desired outcomes, and any accompanying screenshots.

As a member of the GitHub community, your participation is essential. While we can't promise that every suggestion will be implemented, we want to emphasize that your feedback is instrumental in guiding our decisions and priorities.

Thank you once again for your contribution to making GitHub even better! We're grateful for your ongoing support and collaboration in shaping the future of our platform. ⭐

[Aryan-Gore]

This is a valid product-feedback concern focused on cost transparency and administrator control over model fallback behavior.

If GitHub internally switches from the user-selected model to a higher-cost fallback model, customers should have clear visibility into that substitution and predictable control over the resulting usage charges.

The most practical improvements would be:

Show the actual model used in usage and billing records.
Prevent fallback to a more expensive model than the one explicitly selected, unless the user or administrator permits it.
Allow organization/enterprise administrators to disable or restrict fallback models.
Provide a notification or configuration option when fallback could increase AI-credit consumption.
Apply the selected model's credit rate when the higher-cost fallback is mandatory and cannot be disabled.

For the specific incident involving approximately 1,500 AI credits, opening a support ticket is appropriate because the billing record can be investigated against the actual model execution logs.

Overall, the key product request is: model fallback should not undermine the cost controls implied by explicit model selection and organizational policies.

[kosetei]

Thank you for your support. I especially agree that when a higher-cost fallback is mandatory and cannot be disabled, billing should be based on the explicitly selected model's rate.