The era of unlimited AI assistance at a fixed price may be drawing to a close, as GitHub takes the lead in implementing variable costs for its services.
Following a recent pause on new subscriptions for its fixed-price Copilot Pro and Pro+ tiers, GitHub has revealed plans to adopt a consumption-driven billing system effective June.
Subscription fees will stay unchanged, with Copilot Pro at $10 monthly, Pro+ at $39 monthly, Business at $19 per user monthly, and Enterprise at $39 per user monthly.
From June 1, the system will move away from a predefined quota of premium request units—calculated from the volume of AI queries and the capabilities of the underlying models—to AI credits determined by the precise number of tokens processed in interactions.
For instance, Copilot Pro subscribers will continue paying $10 per month but receive credits equivalent to that value instead of a fixed PRU allowance, while Pro+ members will get $39 in credits each month. Business and Enterprise plans will follow a comparable credit allocation based on their fees.
Routine functions such as code suggestions and standard AI operations will not draw from these credits, but more sophisticated tasks like automated code analysis will, according to GitHub. If credits are depleted mid-month, subscribers can purchase additional amounts.
In an official announcement on its blog, GitHub highlighted that the existing premium request unit approach treats a brief inquiry the same as an extended independent programming task in terms of cost, noting that the company has shouldered a significant portion of the rising computational expenses involved.
GitHub stated that this premium request framework has become untenable moving forward.
Ultimately, this represents the conclusion of straightforward fixed-price AI access for GitHub's audience, transitioning to a model tied to token consumption that aligns better with true expenses—though it may impose higher effective costs compared to prior units.
GitHub's adoption of metered pricing could foreshadow similar adjustments across the AI sector for users on unlimited plans.
In reality, the consistent billing options from providers like Anthropic, Google, and OpenAI have served as introductory offers to expand their audiences and encourage adoption of AI-enhanced products.
These major players now face challenges from their achievements, especially with the introduction of resource-intensive agent features into personal subscription levels that rapidly exhaust processing limits.
Anthropic has considered removing advanced agent capabilities and their high token demands from the $20 monthly Claude Pro tier, and both Anthropic and rivals OpenAI and Google have discreetly reduced quotas in their fixed plans, leaving users surprised by abrupt limits.
Anthropic's growth leader, Amol Avasare, noted recently that prolonged agent operations were not anticipated when affordable plans like Claude Pro launched, and existing fixed structures—likely akin to GitHub's PRU method—were not designed to accommodate such demands.
Subtly adjusting limits on fixed plans disadvantages existing customers, yet the shift to consumption billing offers equity and clarity at the expense of potentially much higher charges.
One possible compromise, akin to Anthropic's deliberations, involves retaining fixed pricing for basic conversational AI while applying token fees to complex tools like programming aides and integrated workspace features.
Regardless, the age of unrestricted AI subscriptions seems to be winding down, and for GitHub's community, the reality of variable expenses has now set in.
Ben, a veteran technology journalist with over two decades covering consumer tech, now specializes in artificial intelligence's impact on daily life. His work examines cutting-edge language models and their applications in professional and personal settings to help navigate the coming AI era. As Ben observes, artificial intelligence will transform society more rapidly than anticipated, and regular engagement is key to adaptation. He has contributed to PCWorld since 2014, previously reporting on devices from laptops to surveillance systems before establishing the outlet's AI focus. His pieces have also featured in PC Magazine, TIME, Wired, CNET, Men's Fitness, Mobile Magazine, and others. Ben earned a master's in English literature.