# BYOK (External Provider) Still Consuming GitHub Copilot Premium Request Quota

[Kiznaiverr]

🏷️ Discussion Type

Bug

💬 Feature/Topic Area

VS Code

Body

Summary

When using a BYOK model via Sumopod (external provider like openrouter) in GitHub Copilot Chat (VS Code), GitHub Copilot premium requests are still being deducted from my monthly quota. This contradicts the documented behavior.

---

Expected Behavior

According to the official GitHub Changelog (November 20, 2025):

"API calls made through BYOK are billed according to your provider's terms and do not count against GitHub Copilot usage quotas."

Source: https://github.blog/changelog/2025-11-20-enterprise-bring-your-own-key-byok-for-github-copilot-is-now-in-public-preview/

When sending a prompt through a BYOK model (OpenRouter), the expected outcome is:

• OpenRouter credits are deducted (correct, expected)
• GitHub Copilot premium request quota is NOT deducted

---

Actual Behavior

Both are being deducted simultaneously:

• Sumopod credits are deducted (expected)
• GitHub Copilot premium request quota is ALSO deducted (not expected)

---

Steps to Reproduce

1. Open VS Code with GitHub Copilot extension installed
2. Go to Copilot Chat > Manage Models > Add Models > Custom Endpoint
3. Enter a valid Sumopod API key
4. Select an Sumopod model from the model dropdown in chat
5. Send a prompt in Copilot Chat
6. Check GitHub billing at github.com/settings/billing/summary
7. Check Sumopod usage

Result: Both Sumopod credits and Copilot premium requests are consumed for the same request.

---

Environment

• VS Code version: latest stable
• GitHub Copilot extension version: latest stable
• GitHub Copilot plan: Students
• Sumopod model used: Deepsek Flash V4
• OS: [Windows 11]

---

Evidence

• Sumopod activity log shows the request was processed and credits were deducted
• GitHub billing dashboard at github.com/settings/billing/summary shows premium requests were also deducted at the same timestamp
• This problem occurred after I first installed a skill ui ux pro max. I had never installed any skills before.

---

Additional Notes

This issue may be related to the fact that VS Code occasionally falls back to a Copilot-hosted model without visible indication in the UI. However, in the cases I observed, the model selector clearly showed the OpenRouter model was active at the time of the request.

If BYOK requests are intentionally counted against Copilot quotas for certain plans (e.g. Pro vs Business), this should be clearly documented, as the current documentation does not make this distinction.

Any clarification on whether this is a known bug, a plan-specific limitation, or expected behavior would be appreciated.

[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. ⭐

[roeloffs]

I am seeing the same issue.

I've been using Agent Debug Logs and a custom spreadsheet to calculate my copilot costs for sessions before the June 1st update.

Now seeing that my BYOK session is also consuming AIC, i've done some calculations with my spreadsheet. For a single prompt:
Total cost on the prompt through OpenRouter based off tokens would be around €1.19 (used Minimax M3).
Actual costs according to OpenRouter dashboard: €0.99
Consumed Copilot Credits: 63.20

Meaning that this prompt's total cost was €1.62 instead of €1.19.
So because Copilot decided to still (partially) uses their own model for some reason instead of the chosen OpenRouter model, i have to pay more than 36% extra?

Maybe these numbers are slightly wrong, i have only done a single test but even worse;
Does this mean we cant even use BYOK models if we're out of Copilot Credits?

[Kiznaiverr]

still consuming credit until now, I think I know the problem, from this log Haiku is the model I tried on June 1st before I decided to use BYOK, after that I never used the model from Copilot but instead used the BYOK model, but from the log it looks like there is a code review model, it runs automatically in the background, and eats up Copilot credits

[KaiXuanSama]

Yes, I also encountered this issue last night. While using BYOK for assisted coding, I noticed that my quota was still being consumed. I realized it might be the runSubagent in the agent—when the sub-agent executes tasks, it calls the copilot's own model instead of the BYOK model. The solution is to uncheck "agent" in the configuration tool.

是的，我昨晚也遇到了这个问题，在我使用BYOK进行辅助编码时发现仍然在消耗我的额度，我意识到可能是agent中的runSubagent，在子智能体执行任务时，会调用copilot自身的模型，而不会调用BYOK模型，解决方法就是在配置工具中将“agent”取消勾选

[Kiznaiverr]

update, problem found, copilot runs multiple models, haiku model is used in the background, even though I use BYOK deepseek v4 flash model, and the evidence clearly says 4.3 credits which should not be used for credits if using BYOK

[bogdan-coman-uv]

You can configure the Explore agent to use your BYOK model. There is a setting: github.copilot.chat.exploreAgent.model
If that doesn't work you can copy the agent's .md file and create a custom agent. Replace 'Calude Haiku' with your own model in the frontmatter.