GitHub's AI coding assistant transitioned all plans to a usage-based billing model on June 1, 2026, replacing the previous flat-rate request-based system. The change introduces a new currency — GitHub AI Credits — and makes the cost of each Copilot interaction variable depending on which model is used and how many tokens are consumed.
How AI Credits work
One GitHub AI Credit equals $0.01. Each Copilot interaction draws from a monthly credit balance, with the cost determined by the model tier and token count. Lightweight models used for autocomplete cost fewer credits; premium models used for complex reasoning tasks cost more. All paid plans include a monthly credit allocation, with the option to purchase additional credits.
Credit allocation by plan
Pro ($10/mo): 1,500 credits included. Pro+ ($39/mo): 7,000 credits included. Max ($99/mo): 20,000 credits included. Business and Enterprise customers receive extra credits during the transition period through September 1, 2026.
What this replaces
The previous system charged a flat fee per "premium request" regardless of model or token count. The new system eliminates the premium request definition entirely. GitHub argues that usage-based billing better reflects actual compute costs and lets users mix models within a single session. Critics note that heavy users of premium models could face higher-than-expected monthly bills if usage is not monitored.
Why it matters
GitHub Copilot is the largest AI coding assistant by active users. The billing change affects millions of developers and tens of thousands of organizations. The shift to credits mirrors token-based pricing from OpenAI and Anthropic — signaling that flat-rate AI product pricing is giving way to metered infrastructure economics across the industry.