GitHub Copilot switches to usage-based AI billing
On June 1, 2026, GitHub moved all Copilot plans to usage-based billing, replacing premium request units with GitHub AI Credits consumed by token usage, the company said in its changelog and a blog post. GitHub says 1 AI credit equals $0.01, and that monthly included credits vary by plan, with Copilot Business including $19 and Pro+ and Enterprise including $39 in monthly credits. Reporting from Ars Technica, TechCrunch, The Register, and others describes developer backlash, with some users posting screenshots projecting overage bills from hundreds to thousands of dollars after burning through allotments quickly. GitHub notes those overages apply only if a user sets an additional spending budget; leaving it at $0 stops Copilot rather than charging beyond the subscription. The changelog also adds user-level budget controls, Copilot Max upgrades, and a pause on new sign-ups. Metered AI pricing typically forces teams to add monitoring, budgets, and efficiency work.
What changed
On June 1, 2026, GitHub transitioned all Copilot plans from request-based subscriptions to usage-based billing, the company said in its changelog and a company blog post. Premium request units are replaced by GitHub AI Credits, consumed based on token usage, including input, output, and cached tokens, at published per-model API rates. GitHub says 1 AI credit equals $0.01, so a $10 budget covers 1,000 credits.
Plans and guardrails
GitHub says included monthly credits vary by plan, with Copilot Business including $19 and Copilot Pro+ and Enterprise including $39 in monthly AI Credits. The changelog also adds user-level budget controls, Copilot Max upgrades, code review that consumes GitHub Actions minutes, and a temporary pause on new sign-ups. GitHub notes that additional charges apply only when a user sets an extra spending budget; leaving it at $0 stops Copilot once included credits run out rather than billing beyond the subscription.
The backlash and the takeaway
- •Reporting from Ars Technica, TechCrunch, The Register, and Business Insider describes developer frustration, with users posting screenshots projecting overage bills from hundreds to thousands of dollars after exhausting allotments in hours or days.
- •Generic industry view: moving from flat subscriptions to metered AI billing routinely triggers sticker shock among heavy users and pushes teams to add usage monitoring, budget controls, and prompt- and model-efficiency workflows.
Key Points
- 1Effective June 1, 2026, Copilot uses token-based GitHub AI Credits priced at $0.01 each, with monthly included credits varying by plan ($19 Business; $39 Pro+ and Enterprise), per GitHub.
- 2Developers posted screenshots projecting overage bills from hundreds to thousands of dollars, though GitHub says overages apply only if users opt into an additional spending budget; a $0 budget stops Copilot instead (Ars Technica, TechCrunch).
- 3Editorial analysis: metered AI pricing typically pushes teams toward usage monitoring, budgets, and efficiency work rather than flat-rate planning.
Scoring Rationale
A pricing-model overhaul for GitHub Copilot, one of the most widely used AI coding tools, with immediate, broad developer impact and significant backlash, is a notable products-and-tools story directly relevant to AI/DS/ML practitioners' costs and workflows. It stays in the notable band as a pricing change rather than a capability or platform shift; score unchanged at 6.9.
Sources
Public references used for this report.
View 5 more sources
- 04‘What a joke’: GitHub Copilot’s new token-based billing spurs consternation among devstechcrunch.com
- 05Angry devs vow to flee GitHub Copilot as metered billing takes holdtheregister.com
- 06Devs Sound Off on Usage-Based Copilot Pricing Changevisualstudiomagazine.com
- 07GitHub Copilot Users React To New Usage-Based Pricing Systemnews.slashdot.org
- 08GitHub Copilot users get a rude awakening as new AI pricing goes into effectbusinessinsider.com
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
