GitHub Copilot to Switch to Pay-Per-Use Model
GitHub announced that starting June 1, all Copilot plans will switch to billing based on actual token consumption, with the previous 'Premium Requests' replaced by GitHub AI Credits, and rates aligned with the public API prices of various models.
Code completion and Next Edit Suggestions will not consume credits, and monthly fees remain unchanged (Pro $10, Pro+ $39, Business $19/user, Enterprise $39/user), with additional charges for excess usage at public rates; enterprise customers will have a buffer period from June to August (Business $30, Enterprise $70), and organization-level pooled usage and multi-tier budget controls will be introduced.
In terms of market dynamics, developers and enterprises are shifting to transparent pay-per-use due to high consumption from agentic workflows, with funds moving from fixed packages to high-usage AI coding tools, benefiting providers like GitHub Copilot and models from OpenAI/Anthropic, while costs for users excessively consuming low-priced packages will increase.
Source: Public Information
ABAB AI Insight
GitHub previously suspended new registrations for Pro/Pro+/Student on April 20 and removed Claude Opus, and on April 22 suspended self-registration for Free/Team. This pay-per-use model is the culmination of tightening measures over the past ten days, continuing its response to agentic workflows (sub-agents + long tasks) that lead to per-session costs far exceeding package design.
From a capital perspective, GitHub aligns AI Credits with public API rates, directly passing high consumption costs to users while introducing pooled usage and budget controls. The motivation is to cover costs while keeping subscription fees unchanged, attracting heavy developers for long-term retention and expanding overall ARPU.
Similar to the transition of several AI tool platforms from fixed packages to pay-per-use in 2025 to address the explosive consumption of long reasoning tasks like o1/Claude, GitHub's current adjustment indicates that AI coding tools are at a critical stage of transitioning from low-cost unlimited use to sustainable commercialization billing.
Essentially, this represents a transfer of pricing power: GitHub shifts the pricing power of AI coding costs from fixed subscriptions to actual token consumption, driven by the significant increase in single-user computing power demands from agentic workflows, forcing the platform to implement refined cost control through credits and budget tools, restructuring from subsidized growth to monetization of high-value heavy usage.