"GitHub Copilot is moving to usage-based billing", (github.blog) #198712
Replies: 1 comment
-
|
💬 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 ⏩
Where to look to see what's shipping 👀
What you can do in the meantime 💻
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. ⭐ |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
🏷️ Discussion Type
Bug
💬 Feature/Topic Area
Copilot in GitHub
Body
This is a comment on a recent article from Mario Rodriguez in github.blog Link.
I want to push back on one specific framing in the announcement, because it doesn’t match how Copilot actually worked for heavy users.
The blog says:
“Today, a quick chat question and a multi‑hour autonomous coding session can cost the user the same amount.”
While that’s technically true under PRUs, it’s not the real/whole story.
Under the PRU system, heavy usage wasn’t “free” - it was recognized by how quickly you burned through your premium requests. The model scaled naturally with coding habits because the rate of PRU consumption tracked the intensity of your work. A long coding session didn't JUST cost less; it cost more in PRUs instead of tokens.
The issue wasn’t that users were getting multi‑hour sessions for the price of a chat. The issue was that PRUs were tied to request count instead of the underlying compute cost of modern agentic workflows. That’s a technical limitation of the unit, not a fairness problem with user behavior.
The new token‑based model is probably more accurate from a compute‑billing perspective, but the messaging makes it sound like users were exploiting a loophole. Most of us were just coding, coding requires multi-iteration steps which aligns well with how the agents solve problem; their output are not always correct at first shot — and this is why the old system was predictable in a way developers value, including myself.
I hope the team designing this pricing model includes people who do heavy, real‑world programming work, because the shift from “predictable request units” to “opaque token burn” is a big change in how we plan and budget our workflow. Couldn't get anything done under the new price logic without being largely in debt to billing, please fix.
-- Keny
Beta Was this translation helpful? Give feedback.
All reactions