GitHub will start charging Copilot users based on their actual AI usage (arstechnica.com)

GitHub has announced a shift to a usage-based billing model for its Copilot AI service, effective June 1. This change aims to better align pricing with actual usage amidst increasing demand for limited AI computing resources. Under the new system, subscribers will receive monthly allotments of “AI Credits” corresponding to their subscription fees. Any additional usage will be charged based on token consumption, with costs varying significantly between different AI models, such as OpenAI’s GPT variants which range from $4.50 to $30 per million output tokens. This strategic adjustment is significant for the AI/ML community as it reflects the growing challenge of managing the cost of sophisticated AI services in a rapidly expanding market. By transitioning to a more granular billing approach, GitHub acknowledges the varying computational costs associated with different tasks, such as a simple query versus an extensive coding session. While basic features like code completion will remain accessible without consuming AI credits, advanced functionalities such as code reviews will incur additional costs, prompting users to be more mindful of their utilization and promoting a more sustainable model for AI resource management.