On April 27, 2026, GitHub Chief Product Officer Mario Rodriguez announced that every Copilot plan moves to usage-based billing on June 1. The premium requests system retires, replaced by AI Credits — usage measured in input, output, and cached tokens, billed at each model’s published API rates.
This is more than a change of invoice format. Copilot is the most widely deployed AI coding assistant, and the shift formally ends the flat-rate era of AI help: every chat question and every agent session now carries an explicit price.
The New Billing Model From June 1
The core rule is simple: 1 AI credit equals $0.01. Model interactions are priced from a per-million-token rate table and converted into credits deducted from your plan’s included allowance. Plan prices stay put, and each plan now includes credits equal to its monthly fee: Copilot Pro at $10/month includes $10 in credits, Pro+ at $39 includes $39, Business at $19/user/month includes $19, and Enterprise at $39/user/month includes $39.
Code completions and next-edit suggestions remain unbilled and unlimited on all paid plans. Copilot code review is billed on two tracks: token usage in credits, plus GitHub Actions minutes at standard rates — and the model performing the review is chosen automatically and not disclosed. The docs also specify long-context surcharges above an input threshold: GPT-5.5 doubles from $5 input / $30 output to $10 / $45 per million tokens once input exceeds 272K.
GitHub’s Official Reasoning
The announcement names the contradiction at the heart of the old system: Copilot has evolved “from an in-editor assistant into an agentic platform,” and under premium requests “a quick chat question and a multi-hour autonomous coding session can cost the user the same amount” — a model GitHub calls “no longer sustainable.” Credits make expensive workloads pay expensive prices and lightweight questions pay lightweight ones.
GitHub is also shipping a preview bill experience in early May inside the Billing Overview page, so individuals and org admins can project costs and simulate a typical session’s real spend before the switch lands.
Transition Rules and the Enterprise Grace Period
Monthly subscribers migrate automatically on June 1. Annual Pro and Pro+ subscribers keep premium-request pricing until expiration, but model multipliers rise on June 1 for annual subscribers only; at expiration they drop to Copilot Free unless they upgrade, and converting early earns prorated credit for remaining contract value.
Enterprises get a three-month cushion: through June, July, and August, Business includes $30/month in credits per user and Enterprise $70. Included usage can be pooled across the organization so allowance doesn’t strand with idle seats; admins set budgets at enterprise, cost-center, and user levels, and each org chooses whether overage bills at published rates or hard-caps spending. Fallback models are gone — when credits run out, they run out, and budget controls take over.
Timing matters here: only the prior week, GitHub had tightened limits on Individual plans and paused self-serve Business purchases, framing both as reliability measures. The announcement now promises, “We will loosen usage limits once usage-based billing is in effect.”
Developer Reaction
Visual Studio Magazine’s same-day coverage ran under a blunt headline: “you will get less but pay the same price.” Community pushback clusters around two arguments. First, since included credits equal the plan price, heavy users effectively get less usable work for the same money. Second, agent workflow costs shift from predictable to token-dependent — a misconfigured automation loop can now burn through a budget as fast as it can retry. Subscribers who signed up through the mobile apps face one more restriction: they cannot buy additional credits at all.
What It Means for Teams
Three things are worth doing now. First, when the May bill preview lands, run a full typical session from your main project through it and estimate real credit consumption from data, not vendor arithmetic. Second, revisit model selection: the same task can cost an order of magnitude more credits on a flagship model than a cheap one, which turns model routing from an optimization into a requirement. Third, encode budgets into your process: multi-level admin budgets are already available, and cap modes are the right default for overnight automation and agent jobs inside CI.
The wider context is unit-cost pressure. The same week, DeepSeek shipped a V4 preview claiming frontier parity, pushing open-source cost floors lower still. GitHub’s move to metered billing both exposes the real cost of agentic coding and hands the model-selection lever — along with the invoice responsibility — back to developers.
Sources
- GitHub Copilot is moving to usage-based billing — GitHub Blog
- Models and pricing for GitHub Copilot — GitHub Docs
- Devs Sound Off on Usage-Based Copilot Pricing Change — Visual Studio Magazine
AI-assisted summary compiled from the sources above, reviewed by a human before publishing.