GitHub Copilot shifts to usage-based billing starting June 1

GitHub announced a shift to usage-based billing for its Copilot AI service, effective June 1, aiming to align pricing with actual usage and ensure financial sustainability amid surging demand for AI computing. Currently, subscribers receive monthly allocations of "requests" and "premium requests" that cover diverse AI tasks with varying backend costs. GitHub noted that a quick chat and a multi-hour coding session currently cost the same, and absorbing escalating inference costs is no longer sustainable. Under the new system, subscribers get a monthly allotment of "AI Credits" matching their subscription payment. Additional usage beyond credits will be billed based on token consumption (input, output, cached) using listed API rates per model. API rates vary by model sophistication; for example, OpenAI's GPT-5.4 Mini costs $4.50 per million output tokens, while GPT-5.5 costs $30 per million output tokens. Token usage per prompt also varies with model "thinking" time. Simple AI suggestions like code completion and Next Edit remain free, but Copilot code reviews will incur additional costs via GitHub Actions minutes.