June 30, 2026 marked the close of GitHub Copilot's first complete 30-day billing cycle under usage-based pricing — and for many developers, the invoice was a shock. Since GitHub moved every Copilot plan to token-metered "AI Credits" on June 1, agentic users have reported monthly costs running 10x to 50x higher than the old flat-rate plans, with projections circulating online of bills jumping from $29 to $750 and from $50 to $3,000.
Key Highlights
- All GitHub Copilot plans switched to usage-based AI Credits billing on June 1, 2026, replacing the older premium-request model.
- One AI Credit equals $0.01, and Copilot now bills against the input, output, and cached tokens each request consumes at published per-model API rates.
- Pro costs $10/month and includes $15 in credits; Pro+ is $39/month with $70 included; Max is $100/month with $200 included. Business and Enterprise seats draw from a shared, organization-pooled balance.
- Code completions and Next Edit suggestions remain free across all plans.
Details
The mechanics are simple but unforgiving. Instead of counting discrete "premium requests," Copilot now meters raw token consumption — every input prompt, every generated output, and even cached context — and converts it into credits at the API rate of whichever model handled the work. A one-line chat question and a multi-hour autonomous coding session are no longer treated as economically equivalent.
That distinction hits agent-mode users hardest. One developer reported watching 82 percent of their monthly credit allowance disappear on the first day. Another ran a single agentic session that cost $38 — nearly four times their entire Pro plan budget. On Reddit and GitHub's own community forums, threads filled with annotated screenshots and cost projections. TechCrunch captured the mood in a May 30 headline quoting a developer: "What a joke."
Spending controls exist, but they are opt-in. Organizations can configure budget limits at the enterprise, cost-center, and user levels, then choose whether to block further usage or allow overage billed at published per-token rates once the included pool is depleted. Without that configuration, there is no automatic ceiling.
Impact
For solo developers and small teams, the change converts a predictable monthly line item into a usage meter that can spike without warning. The heaviest agentic workflows — long-running autonomous tasks, repository-scale context, and frontier-model code review — are precisely the features GitHub has been promoting, and they are also the most expensive to run.
For engineering managers, the takeaway is operational: AI coding spend now behaves like cloud compute, demanding budgets, alerts, and per-team caps rather than a fixed seat license. Teams that adopted Copilot expecting flat costs must now treat it as a variable expense to forecast and govern.
Background
GitHub framed the shift as a matter of sustainability. The company argues Copilot "is not the same product it was a year ago," having grown from an autocomplete tool into an agentic platform spanning chat, code review, repository-scale context, agent mode, cloud agents, and higher-end frontier models. Under that load, GitHub says the old flat model was "no longer sustainable," and that usage-based pricing "better aligns pricing with actual usage."
The transition has been contentious since it was first announced in late May. Coverage from TechCrunch and Visual Studio Magazine documented early developer pushback well before the first full cycle closed, and the June 30 invoices have now turned projections into real numbers.
What's Next
Expect the debate over predictable versus consumption-based AI pricing to intensify across the developer-tools market, where rivals like Cursor and Anthropic's Claude Code are navigating the same tension between flat subscriptions and metered token costs. For Copilot users, the immediate priorities are setting spending caps, auditing which workflows burn the most credits, and deciding whether agent-heavy usage justifies a higher tier — or a switch.
For developers across the MENA region, where engineering budgets are often tightly managed against fluctuating exchange rates, the message is sharper still: AI assistant costs are now a variable to monitor, not a fixed subscription to forget. Configuring spending limits before the next cycle begins is no longer optional.
Source: TechCrunch