GitHub Copilot has moved from a flat subscription to token-based billing, charging developers for every token used in AI interactions. This shift has sparked immediate backlash from 4.7 million subscribers who now face unexpected costs beyond their monthly fees.
- GitHub replaced flat-rate plans with AI Credits costing $0.01 each, consumed per token used in inputs, outputs, and cached context.
- Subscription tiers remain priced at $10, $19, $39 per month, but these now represent monthly credit allowances, not spending limits.
- Hitting monthly credit caps triggers hidden spending limits that block additional purchases unless users upgrade tiers.
- Annual subscribers face accelerated credit usage due to multiplier increases, draining quotas up to four times faster.
From flat fees to token credits
As of June 1, GitHub Copilot’s pricing model shifted to token-based billing using AI Credits priced at $0.01 each. While subscription fees for Copilot Pro, Pro+, Business, and Enterprise plans remain unchanged, these payments now grant monthly credit allowances rather than unlimited usage.
Developers consume credits for AI interactions beyond basic code completions and Next Edit Suggestions, which remain free and unlimited. This token metering includes inputs, outputs, and cached context tokens, making usage more granular and potentially costly.
Hidden spending caps frustrate developers
Two crucial yet underemphasized details have fueled developer dissatisfaction. When users reach their monthly credit limits, GitHub imposes temporary spending caps based on account age and verification status. Instead of allowing pay-as-you-go top-ups, affected users must upgrade to higher subscription tiers to regain premium features within the billing cycle.
Additionally, annual subscribers on legacy plans have encountered a multiplier change increasing token consumption rates significantly. For instance, Claude Opus 4.7 requests now cost 27 times more tokens compared to 7.5 times previously, accelerating quota depletion, while GPT-5.4 usage increased sixfold. Despite no direct token bills yet, these changes effectively reduce usable request volumes.
Real-world impact on developer costs
Data shared by developers reveals sharp increases in monthly expenses. An enterprise team that previously billed around $500.35 monthly now faces estimated costs near $5,290.92 under token billing. On the $10 Pro tier, a 20 to 30-minute coding session consumed 16 percent of a user’s entire monthly credit allowance, highlighting how quickly credits can deplete during typical workflows.
Pooling credits eases burden for organizations
Business and Enterprise subscribers benefit from pooled credits at the organizational level. This system automatically balances unused credits from low-usage members against heavy consumption by developers engaged in intensive AI sessions. To mitigate the transition, GitHub is also providing promotional credit buffers through August, with $30 extra per user for Business plans and $70 for Enterprise.
Despite these measures, the shift reflects growing costs for AI-powered tools. Internal Microsoft documents indicated Copilot’s operating expenses nearly doubled week-on-week earlier this year, necessitating the pricing overhaul. The challenge remains balancing sustainable economics with developer expectations and usability.
Via: Notebookcheck
