A new billing model allows you to work only 2 days a month.

[mqhe2007]

🏷️ Discussion Type

Product Feedback

💬 Feature/Topic Area

Copilot Agent Mode

Body

That's NB！ NB = niubi = awesome

[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. ⭐

[MohabAnwar921]

Yup, burned half my quota to apply one change in one script

[dbestechno]

3 Days In. Already Past 60%.
Three days. Not a sprint, not a hackathon, not a stress-test. Just regular, everyday use — and the meter is already past the halfway mark, racing toward empty like it was designed to run out.
Whoever built this new billing plan clearly tested it in a vacuum. The limits feel arbitrary. The usage tracking is opaque. And the upgrade nudges? Already showing up on day two.
This isn't a pricing model. It's a pressure campaign.
A plan that can't survive a Tuesday is not a plan. It's a trap.
Normal use should not feel like rationing. If you're anxious about every action you take within a product you're already paying for, the product has failed you. Full stop.
Three days in, over 60% consumed, twenty-seven days still to go. Someone designed this. Someone approved this. And someone needs to fix it.

[DarioTrima81]

The new Copilot credit system is a complete mess!
it's making the AI ​​completely unusable despite using caveman, careful prompt sizing, code-graph, etc.

I'm asking the AI ​​for help debugging a problem running a C applet in RAM on a Cortex M0.
Today I asked 3 questions and reached 25% of my available tokens for the entire month!!!!! 25% with 3 damn questions.

I went from 3% -> 25% by asking an auto model agent (not even Sonnet...a simple auto model):
"the change I made doesn't work, analyze memory.md and the code changes, the terminal output is: SAM-BA: v2.1 [build:Jun 1 2026 12:26:16]
Written 320 B to 0x20000900
GO_ACK_FAIL TimeoutError("Timeout waiting marker b'G\n\r'; got b'A\n\r'")
Sniffing 3.0s...
RX_RAW
RESULT_READ_FAIL SerialTimeoutException('Write timeout')."

And the AI had to checkout only 3 files, one .md (70 rows), one .c (1312 rows) and one .py (104 rows).

It's not like I'm using the AI ​​to draw rainbow-colored mini-ponies flying in the blue sky, or asking it to write me down and entire Linux Kernel from scratch...in this case, just debugging a small C applet running in RAM.

I'm using the AI to help me solve work problems that I can't solve on my own, by paying for a service, but the changes that have been made to the payment system translate into a robbery and a total devaluation of the prepaid credit.

This is unacceptable!

[TroyCoderBoy]

I heard 4 days working weeks are becoming a thing in some countries, but it looks like Github is taking it to the extreme with 2 day working months.

[mqhe2007]

The copilot server room administrator can start taking a vacation tomorrow, as the machines can now be shut down.

[AGTon26]

this is a mess and a terrible way to handle customers by not even replying.

[whitewidovv]

2 DAYS? oy vey, i should be so lucky.

my $39 plan was completely used up in about 2 HOURS of normal use. I will not let my subscription renew if there isn't a change to this. i'll be finding another provider. i heard codex is good these days, may try that. i used claude previously and didn't hate it.
i just loved the agent system within copilot. but new cost model is unsustainable and I will not be resubscribing.

also, i have never before posted on github community before today; never felt the need to, but this change was so bad i felt the need to come and speak out about it. github, microsoft, i hope this is seen and understood and rectified. i get they were trying to balance things, but they balanced things in the same way anakin skywalker brought balance to the force

[mqhe2007]

You're really unlucky, I've already unsubscribed and am looking for new alternatives, welcome to recommend!

[TroyCoderBoy]

I switched to Claude Code Max. Expensive but a flat monthly fee. I can use it as much as I want all day and never get close to limits.

[AGTon26]

I am heading that way next and this should be made very public on all their social media accounts. i have gotten 2 sessions of use and have a very important project to get finished. I have tried several times to contact support but they never reply and this was before this new setup they have. this is straight BS

[DarioTrima81]

Switch to higher tier is their plan...it reminds me a Black Mirror episode!

At the beginning, they gave us AI free of charge, to push the adoption, to force us integrate the AI in our workflow.

Then they started to ask us a payment (I mean that's fine, no discussion).

Now you have AI in browser, AI in Office, AI in IDE, AI mostly in every laptop...probably there is an AI even in some toilet paper, I don't know.

Finally, now that we are all used to AI and have been forced to use it massively, they started the new mechanincs: reduce the time/functions you can use, increase the cost, reduce again the time/function, increase again the cost, forcing us to move to highr tier.

[aisaya-ana]

Hehe, I also got 2 days to work per month as well XD

[AGTon26]

yes this is a straight mess of a system and needs to me made public, so that they will at least try to hear their customer base. but it seems like they dont care. I have tried contacting them for months to no reply. And now this mess of a credit system has made it 10x worse.

[TroyCoderBoy]

@AGTon26 Looks like it wasn't a money maker for them and this is their way of ditching the service - making it unaffordable.

[AGTon26]

that is what i said as well. I have been trying to contact them for months and no reply. no help whatsoever. This is very unprofessional. we have to all slam their social media accounts and hopefully prevent others from falling to the dame fate. I just need to finish this project of mine and on to openrouter with my cline setup.

[aisaya-ana]

I never try else Agent yet, maybe it time for new Agent. XD

[yunze7373]

NO! Only 30mins

[riaje]

Well, I try to use different models to slow the usage but I should admin 24% for 2 hours of works on a copilot pro + is a bit sad.

[aiis2]

You're lucky; I used up my 7,000 yuan credit limit in 3 hours.

[kendrick-gs]

this is just ridiculous, tot it was 5 hours session limit turns out to be reset next month!?!

[aiis2]

30 day

[eqfwcev123]

Immediately cancelled my plan after trying the new model for 1 hour. Does paying 39USD to use AI for 2 hours even makes sense?

[mqhe2007]

I think Microsoft might really be using this method to make us proactively cancel our subscriptions and then cut off this business line. Otherwise, it's hard to understand the significance of the new model.

[TroyCoderBoy]

Maybe they just grabbed an easy 39USD off everyone before we all walk out the door and they close it all down.

[moonlune]

I think there’s a huge problem with Copilot’s billing model, which charges extra based on connection duration! This is truly a ridiculous calculation method! To factor in everything—model response times, errors, repeated trial-and-error attempts to solve problems, connection issues, and so on—and then pass the cost on to the user? This billing model is downright absurd!
Are they trying to drive all users away?

[TroyCoderBoy]

I regret to report it's working as designed.

[vivekbhat0120]

As of today (June 1st, 2026), I’ve been using the GitHub Copilot extension in VS Code (Pro plan), and I recently switched to the Claude Sonnet 4.6 model in agent mode. I ran approximately 5 to 6 individual prompts, and my entire monthly credit allocation was depleted, despite only using a handful of requests.

What concerns me most is that in the latest update, the model selection interface no longer displays credit consumption per request. Previously, I could see how many credits each model consumed, but now I only see the model name. This lack of visibility is preventing me from effectively tracking and managing my credit usage, and as a result, I unexpectedly exhausted my quota.

I would like to request: (1) clarification on why credit cost per request was removed from the UI, and (2) a timeline for when this will be restored or if there’s another way to monitor usage accurately.

This is critical for budgeting and ensuring that I don’t overrun my credits unexpectedly. I appreciate your urgent attention to this matter.

[sinnitesh]

I also exhausted my pro credits in just two days. Now I’m free to go home and spend more time with my family, thanks for the 'work-life balance' update.

[dbestechno]

3 Days In. Already Past 60%.
Three days. Not a sprint, not a hackathon, not a stress-test. Just regular, everyday use — and the meter is already past the halfway mark, racing toward empty like it was designed to run out.
Whoever built this new billing plan clearly tested it in a vacuum. The limits feel arbitrary. The usage tracking is opaque. And the upgrade nudges? Already showing up on day two.
This isn't a pricing model. It's a pressure campaign.
A plan that can't survive a Tuesday is not a plan. It's a trap.
Normal use should not feel like rationing. If you're anxious about every action you take within a product you're already paying for, the product has failed you. Full stop.
Three days in, over 60% consumed, twenty-seven days still to go. Someone designed this. Someone approved this. And someone needs to fix it.

[marchlqq10]

I used 42% of my allowance. As of yesterday, I have completely abandoned GitHub Copilot; I am now using other AI software instead. It is just as fast and stable, and the resource consumption remains just as low as it was before Copilot's billing model changed. So, what are you all still waiting for?

Hello,

We have provided more details and a FAQ about usage-based billing in this discussion, which may be updated if new information becomes available.

To keep conversations all in one place and to ensure you receive timely updates, we’ll be closing and locking this post and recommend you comment on or subscribe to the discussion linked above. Thank you!