GitHub will begin charging Copilot users based on their AI usage

GitHub has he announced that it will be moving to a payment method for using its GitHub Copilot AI service as of June 1. The move is positioned as a way to “better align pricing with actual usage” and is a necessary step for Copilot to be financially sustainable amid the high demand for limited AI computing resources.

GitHub Copilot subscribers for now receive a monthly portion of “requests” and “value requests,” which is used every time the Copilot asks for help from the AI ​​model. But those broad categories cover many AI tasks with significant computing capital, GitHub says.

“Today, a quick and independent query of a few hours can destroy the same user,” Microsoft Corporation he wrote in his announcement. And even though GitHub says it has “taken the majority of the growing number of applications” so far, combining “all requests” together is “not sustainable.”

Under the new pricing structure, GitHub Copilot subscribers will receive a monthly installment of “AI Credits” equal to their monthly payment. Pricing for AI usage beyond that amount “will be calculated based on how the tokens are used, including input, output, and stored tokens, using the API pricing listed for each type.”

Those API prices can vary greatly depending on the model being used; OpenAI’s lowest GPT version prices various from $4.50 per million token issuance (GPT-5.4 Mini) to $30 per million token issuance (GPT-5.5), for example. The total number of tokens used in an AI event can also vary depending on the “thinking” time that the model needs to generate its output.

GitHub Copilot subscribers will be able to use simple AI suggestions like code completion and Next Edit without consuming AI. But Copilot code comments will come with other costs in the form of GitHub Actions moment.