GitHub Copilot switches to metered token billing June 1 as flat-rate model proves unsustainable

Microsoft's GitHub is ending its flat-rate AI billing model for Copilot, shifting to usage-based token metering on June 1, 2026. The company acknowledged that its current request-based pricing structure has become financially unsustainable.

New pricing structure

Under the new system, GitHub introduces "GitHub AI Credits" as a virtual currency unit worth $0.01 each. Subscription prices remain unchanged, but users now receive monthly credit allocations:

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

Existing Business and Enterprise customers receive higher temporary allocations through September 1, 2026: 3,000 and 7,000 credits respectively.

Why the change

"GitHub has absorbed much of the escalating inference cost behind that usage, but the current premium request model is no longer sustainable," wrote Mario Rodriguez, GitHub's chief product officer. Under the old model, a brief chat question and a multi-hour autonomous coding session cost users the same amount, while complex prompts requiring extended "thinking" time often cost GitHub more than subscription revenue.

Usage will now be calculated based on input tokens, output tokens, and cached tokens, with rates varying by model. GitHub will provide a preview billing experience in early May showing projected costs before the transition.

Industry-wide correction

The shift follows increased AI agent usage and growing developer experimentation with AI coding tools that have strained inference infrastructure across the industry. GitHub last week suspended creation of new Copilot Pro, Pro+, and Student plans. Anthropic and Google have similarly limited service usage, while OpenAI introduced a $100/month subscription tier and is considering ending unlimited usage under existing plans.

Users who exhaust their monthly credits can set overflow budgets or wait for the next billing cycle. Annual subscribers have the option to cancel and receive refunds.

What this means

GitHub's move to metered billing represents the end of the "all-you-can-eat" AI era for developer tools. The non-deterministic nature of token consumption means users cannot precisely predict costs in advance—different prompts may involve tools that complicate calculations. This uncertainty, combined with the need to monitor credit balances, fundamentally changes the economics of AI-assisted coding. Organizations will need to implement usage policies and budget controls, while individual developers may self-limit their AI tool usage to avoid unexpected costs. The three-month grace period for existing enterprise customers suggests GitHub expects significant pushback as users adjust to the new reality of metered AI services.