Microsoft's GitHub is closing the all-you-can-eat AI buffet for Copilot users, moving from unlimited request-based billing to metered token consumption on June 1, 2026. The shift reflects a broader industry reckoning with unsustainable AI inference costs, as companies like Anthropic, Google, and OpenAI have all begun tightening access to their services.

Why Is GitHub Abandoning Unlimited AI Billing?

Under the current request-based system, GitHub absorbs the cost difference between what users pay and what the company actually spends on AI inference. The problem: a quick chat question and a multi-hour autonomous coding session cost the user the same monthly subscription fee, even though the latter consumes vastly more computational resources.

"Today, a quick chat question and a multi-hour autonomous coding session can cost the user the same amount. GitHub has absorbed much of the escalating inference cost behind that usage, but the current premium request model is no longer sustainable," explained Mario Rodriguez, chief product officer on the GitHub product team.

Mario Rodriguez, Chief Product Officer, GitHub

The cost crisis accelerated after February 2026, when OpenClaw attracted widespread attention and sparked a surge of experimentation with AI agents running 24/7 on various tasks. Simultaneously, improving AI model capabilities encouraged more developers to explore AI-powered coding, overwhelming GitHub's infrastructure and forcing the company to acknowledge it could no longer subsidize unlimited usage.

How Will the New Token-Based Pricing Work?

GitHub is replacing request-based billing with usage-based billing tied to token consumption. Tokens are sets of three or four characters that represent the basic economic unit for selling AI services. Instead of counting premium requests, every Copilot plan will include a monthly allotment of GitHub AI Credits, a virtual currency worth $0.01 each.

Users will be charged based on input tokens, output tokens, and cached tokens, with different models metering tokens at different rates. The challenge: usage-based billing is non-deterministic, meaning users cannot know in advance exactly how many tokens a model will consume to respond to a specific prompt. Different prompts may involve tools that complicate token consumption calculations.

To help users navigate this uncertainty, GitHub plans to introduce a preview bill experience in early May, giving users and administrators visibility into projected costs before the June 1 transition.

What Are the New Pricing Tiers and Monthly Allowances?

GitHub's subscription rates remain unchanged, but monthly AI Credit allowances vary by plan:

• Copilot Pro: $10 per month with 1,000 AI Credits per month
• Copilot Pro+: $39 per month with 3,900 AI Credits per month
• Copilot Business: $19 per user per month with 1,900 API Credits per user per month
• Copilot Enterprise: $39 per user per month with 3,900 API Credits per user per month

Existing Copilot Business and Enterprise customers receive a temporary boost: 3,000 and 7,000 API Credits per user per month respectively, from June 1 through September 1, 2026.

Once users exhaust their monthly AI Credit allowance, they can define an overflow budget to continue using premium features, or simply wait until the next billing cycle resets their balance. However, code completions and Next Edit Suggestions remain unlimited on paid plans, even after users reach their credit limit.

How Much More Expensive Will Premium AI Models Become?

The pricing multipliers for premium models will increase significantly under the new system. For example, Anthropic's Opus 4.7 model currently carries a 7.5x multiplier under request-based billing; under token-based billing, that multiplier jumps to 27x. OpenAI's GPT-5.4 will see its multiplier rise from 1x to 6x.

Users on annual subscription plans have two options: cancel and receive a pro-rated refund, or be downgraded to Copilot Free upon subscription expiration. Annual plans will not be renewable under the new system.

Steps to Prepare for the June 1 Transition

• Review Your Current Usage: Check how many premium requests you submit monthly and estimate which new plan tier best matches your needs before the June 1 deadline
• Monitor the Preview Bill: Use GitHub's preview bill experience launching in early May to see projected costs under token-based billing before the transition takes effect
• Set Overflow Budget Limits: Decide in advance whether you want to allow overflow spending once your monthly AI Credits are exhausted, or prefer to pause AI usage until the next cycle
• Evaluate Annual Plans: If you're on an annual subscription, decide whether to cancel for a refund or ride it out until expiration, knowing premium model prices will increase significantly

GitHub's shift reflects a broader industry pattern. Anthropic and Google have already taken steps to limit some uses of their services, while OpenAI debuted a more expensive $100 subscription tier to boost usage of its Codex model. Cloud providers like Amazon Web Services (AWS) and Microsoft Azure have also faced capacity challenges as demand for AI inference outpaced infrastructure.

The move signals that the era of subsidized, unlimited AI access through subscription plans is ending. Companies are shifting the cost burden directly to users through metered billing, forcing developers to make conscious choices about when and how they use AI tools. For GitHub Copilot users, the June 1 transition marks the end of the "all-you-can-eat" model and the beginning of a more cost-conscious era of AI-assisted coding.