GitHub Copilot’s New Token-Based Pricing Sparks Backlash GitHub’s decision to overhaul Copilot pricing from a simple monthly subscription to metered “AI credits” has triggered alarm among many developers who say their costs are suddenly unpredictable and, in some cases, unaffordable.

Early warning signs

Concerns began mounting in late May, when reports detailed that the “golden age” of low, flat-rate Copilot access was ending. GitHub planned to replace its subscription with a token-based billing system that could “bill users at a significantly higher rate,” especially impacting smaller companies and individual developers.

June 1: The switch flips

On June 1, GitHub Copilot officially moved from a request-based model to token-usage billing. Usage is now converted into AI credits, each equal to $0.01, with costs determined by the specific AI model and number of tokens consumed. Internal cost estimators quickly showed some power users burning through monthly credit allotments within a day or two, with projected bills “hundreds of dollars more than previous months.”

Developers began sharing screenshots of estimated charges jumping from under $30 or $50 a month to hundreds or even thousands of dollars, prompting one Redditor to call the new system “stupidly expensive” and “no longer cost-effective or useful in any practical way.” Another user’s projected bill reportedly soared from about $44 to over $750, drawing comparisons to Uber’s early low prices followed by steep hikes once users were dependent.

Community backlash and counterarguments

As the change took effect, coverage highlighted users who “burned through their whole monthly ‘AI credit’ allotment in a single day,” underscoring the shock over how quickly credits disappeared. Some developers argued the backlash reflected unsustainable “vibe coding,” suggesting that efficient, experienced users weren’t seeing such extreme overages.

GitHub, for its part, had warned in April that the old pricing was “no longer sustainable” and that costs needed to align with expensive underlying AI models. Supporters of the shift say usage-based billing more accurately matches what the service costs to run—while critics fear it will price out the “little guy” and push heavy users to rethink how, or whether, they rely on AI coding tools.

Continue reading https://foxvector.com/stories/019e8d30-1ef9-1a74-70d6-07c3b600eb66