News

Devs Sound Off on Usage-Based Copilot Pricing Change: 'You Will Get Less, but Pay the Same Price'

• By David Ramel
• 04/27/2026

GitHub Copilot users are pushing back on GitHub's planned move to usage-based billing, with some developers arguing the change will make the AI coding assistant less predictable and less valuable even though base subscription prices are staying the same.

The change, set for June 1, 2026, replaces premium request units, or PRUs, with GitHub AI Credits. Instead of charging against a request-based allowance, Copilot usage will be calculated by token consumption, including input tokens, output tokens and cached tokens, using published rates for each model.

GitHub detailed the change in an April 27 blog post, saying Copilot has evolved from an in-editor assistant into "an agentic platform capable of running long, multi-step coding sessions, using the latest models, and iterating across entire repositories." The company said a quick chat question and a multi-hour autonomous coding session can currently cost the user the same amount, and that the PRU model "is no longer sustainable."

"Instead of counting premium requests, every Copilot plan will include a monthly allotment of GitHub AI Credits, with the option for paid plans to purchase additional usage," GitHub said. "Usage will be calculated based on token consumption, including input, output, and cached tokens, using the listed API rates for each model."

In an April 27 email to subscribers, GitHub said: "To keep up with the way you use Copilot, we're transitioning to usage-based billing, and we want to give you enough time to prepare. If you're also an admin on a Copilot Business or Enterprise plan, you'll receive a separate email covering what's changing for your organization." The email also said: "Last week, we made temporary updates to Copilot individual plans to improve reliability and performance ahead of the broader move to usage-based billing. We will loosen usage limits once usage-based billing is in effect."

A related GitHub Community FAQ and discussion thread drew immediate reaction from users focused on token costs, Opus model access, rollover questions, refunds and the loss of request-based predictability. At the time of this writing, there were 70 comments and 105 replies, but that FAQ appears to have been published April 17, while the blog post announcement was just published today, so that is guaranteed to change. The number of comments is actually climbing while I write this. See some of the comments below.

Base Prices Stay the Same, Billing Unit Changes
GitHub said Copilot Pro will remain $10 per month and include $10 in monthly AI Credits, while Copilot Pro+ will remain $39 per month and include $39 in monthly AI Credits. Copilot Business remains $19 per user per month and Copilot Enterprise remains $39 per user per month, with corresponding monthly AI Credits included.

For Business and Enterprise customers, GitHub said it is also introducing pooled included usage across a business. That means unused credits are not isolated to each seat, allowing lighter users' unused included usage to offset heavier users in the same organization.

GitHub said code completions and Next Edit Suggestions will remain included in all plans and will not consume AI Credits. Other AI-model-based Copilot features, including agentic workflows and code review, will consume credits. Copilot code review will also consume GitHub Actions minutes in addition to AI Credits.

Annual Pro and Pro+ customers will remain on premium request-based pricing until their plan expires, GitHub said, but model multipliers will increase on June 1. At expiration, annual-plan users will transition to Copilot Free with the option to upgrade to a paid monthly plan. GitHub also said annual-plan users may convert to a monthly paid plan before expiration and receive prorated credits for the remaining value of their annual plan.

Likely Effect: More Cost Management for Agentic Use
The likely effect depends heavily on how developers use Copilot. Users who mainly rely on inline code completions and Next Edit Suggestions may see limited day-to-day impact because those features remain included. Users who rely on Copilot Chat, large context windows, code review, agents or long-running coding sessions are more likely to see usage limits and overage decisions become part of their workflow.

The move also shifts the user's mental model from counting requests to managing token consumption. Under the current system, a request is counted against a PRU allowance. Under the new system, the same prompt can cost different amounts depending on model choice, context size, output length, cached-token use and any intermediate tool activity in an agentic workflow.

That creates a potential tradeoff for GitHub and its customers. For GitHub, the change aligns revenue more closely with model inference costs. For developers, it makes Copilot resemble a metered cloud service, where usage dashboards, model selection and budget controls become part of routine cost management.

Users Question Predictability and Value
The GitHub Community thread reflected broad concern that the shift could reduce the value of paid Copilot subscriptions even if list prices do not change. Several users said request-based billing was easier to understand because developers could predict how much work they could get from a plan before hitting limits.

One user said the difference between request-based and token-based billing "could sober up an alcoholic from mere shock," adding: "You will get less, but pay the same price."

Another user said token-based billing makes usage harder to predict because a request's cost depends on the prompt, tools, files, model and output. "Previously, one request (prompt) was calculated as one - regardless of the details involved in producing the result," the user wrote. "In essence, the irony is that your 'Usage-based billing' will make it significantly harder for you to predict your.. usage. How much you will need or use is dynamic, heavily context-dependent, and non-deterministic."

Another commenter framed the issue as a product-design problem: "Request-based billing is user oriented. Users do not care about how the AI works, they only care about the input and the output."

Some users focused on whether Copilot remains compelling if it becomes closer to direct API consumption. One commenter wrote: "There's just one thing I don't understand: if I have to pay per token, where's the advantage compared to using the API of my favorite model(s) directly from the production environment?"

Another said the change could make Copilot "effectively becoming a wrapper over API consumption" while not giving users the same controls they would have when using an API directly, such as max tokens, temperature or system prompts.

Sample User Comments from the FAQ Thread
Representative comments from the GitHub Community discussion are listed below. GitHub obviously anticipated some of these concerns, as one of the questions it listed in its FAQ read: "This just wiped GitHub's value moat -- why should I stay?" GitHub's answer to its own question was: "We believe GitHub Copilot remains the best value and experience for agentic coding. Usage-based billing aligns cost more closely to actual usage and value, while continuing to offer developers the freedom to choose the models and agents that work best for them."

• "Request-based billing is user oriented ... Token-based billing is not user oriented."
• "This is removing the one real advantage GHCP had over Claude Code et al."
• "I would much rather you jacked up the price per premium request by 30-50x than switch to token based pricing where agents and so forth can nuke your supply of credits in very little time."
• "Can we please get better insights and metrics on token usage?"
• "Unless the credits rollover, then this plan does not 'align pricing to actual usage and costs'"
• "Removing the free models destroys any value in the individual plans"
• "I might just unsubscribe because of this very reason!"
• "What users need right now to make an informed decision is a clear cost per token."
• "Saying 'plan pricing isn't changing' while altering what's included in those plans is misleading without that level of transparency."
• "bye bye Github Copilot Pro"
• "AI credits"?! More number games."
• "I'm done!"
• "Great way to lose the one thing that set you apart from the rest of the industry."
• "Joke aside, this 'price increase' is huge."
• "I have canceled my subscription."
• "At this point, I think it makes more sense to use something like kilo.ai and get access to 500+ models on OpenRouter via a similar VS Code extension or CLI."
• "This feedback is not coming from excessive or abusive usage patterns."
• "Copilot has been a strong product. The concern is that, under this model, it risks losing the advantages that made it compelling in the first place."
• "is it even worth it to keep subscribing to pro when we only 10$ monthly credits?"
• "It was THE reason I was staying on GitHub."
• "Time to move on elsewhere."
• "Today is a sad day .. it's obvious that GitHub isn't a charity .. but the way you're handling this change is horrible"
• "RIP GHCP"

Model Access and Rollover Questions
Several comments focused on Anthropic's Claude Opus models. Users asked whether GitHub would restore Opus access if Copilot usage is being billed by tokens. One commenter asked: "Okay, with this change, will you at least bring back Opus since we're all paying for this again?" Another asked: "At least bring back Opus 4.6? 4.5?"

Another user asked whether Copilot Pro users would regain access to Claude Opus 4.6 or another high-effort Opus model once usage-based billing starts, arguing that if usage is metered by tokens and AI Credits, GitHub should explain why Opus remains tied to Pro+.

Rollover was another recurring topic. Users asked whether unused AI Credits will carry over to future months or expire at the end of each billing period. One commenter wrote: "if it is a usage based i dont want my money to just burn by the end of the month."

Another said: "Maybe we're talking of a few dollars, but then again those are plans targeted at individuals, so it could matter. Nonetheless, given that we are now speaking of "usage-based billing", for completeness one should explain what happens when I use less tokens than the allowed ones."

Refunds and Cancellations Enter the Discussion
The discussion also included users saying they had canceled or were considering canceling. One user asked whether a refund was available for the remainder of an annual plan, writing: "I've stopped using Microsoft GitHub Copilot altogether."

A GitHub Community Admin replied that a refund can be requested for an individual plan through Settings, Billing and licensing, Licensing, Manage subscription, and then Cancel and refund "subscription." The admin said the option will be available until May 20.

Other commenters said the change makes alternatives more attractive. One said: "At this point, I think it makes more sense to use something like kilo.ai and get access to 500+ models on OpenRouter via a similar VS Code extension or CLI." Another wrote: "Hello all, I already canceled my subscription. Good luck, Github Copilot."

GitHub's Balancing Act
GitHub's case for the change is that agentic coding creates higher and more variable compute costs than earlier Copilot workflows. The company said the current model requires GitHub to absorb escalating inference costs and does not distinguish between a brief chat question and a multi-hour autonomous coding task.

The FAQ thread shows the user-side issue: many developers valued Copilot because the request-based model made usage relatively predictable. Moving to AI Credits may give GitHub a billing model aligned with compute costs, but users are asking for clearer token-cost visibility, better usage metrics, rollover answers and model-access policies that match the new usage-based structure.

The result is a pricing change that may not affect every Copilot user equally. Completion-heavy users may see little difference. Developers using Copilot as an agentic coding environment may face a new calculation: how much work a $10 or $39 monthly credit allotment actually buys, and whether Copilot remains the best place to spend that AI budget.