GitHub Copilot kills free fallback models as token-based billing goes live June 1.
GitHub confirmed April 27 that every Copilot plan will switch from premium request units to token-based billing on June 1, replacing predictable monthly allowances with consumption metered by the exact model used, input and output volume, and cached tokens.
The headline: plan prices stay flat (Pro at $10, Pro+ at $39, Business at $19, Enterprise at $39), but what those dollars buy is shrinking. Each plan now includes AI Credits equal to its subscription cost, $10 worth of credits for Pro, $39 for Pro+.
Once those credits are gone, there's no safety net.
GitHub is eliminating the free fallback model that previously let users who exhausted their premium requests downshift to a lower-cost model (GPT-5 mini, GPT-4.1, or GPT-4o) and keep working. Under the new system, hitting your credit limit means buying more credits or stopping entirely. As Heise Online reports, "A fallback to a free, simple model as before will no longer be available." The change stems from Copilot's evolution into what GitHub calls "an agentic platform capable of running long, multi-step coding sessions." A quick chat question and a multi-hour autonomous coding session currently cost users the same amount under request-based pricing. GitHub said in its announcement blog post that the old model "is no longer sustainable."
Per-token costs vary dramatically by model. One million output tokens runs from $1.25 for GPT-5.4 nano to $30 for GPT-5.5.
Claude Opus 4.7 costs $25 per million output tokens; Gemini 3.1 Pro costs $12. A single agentic session using a powerful model can burn through a Pro user's entire $10 monthly credit allotment quickly.
Annual plan subscribers face an additional squeeze: model multipliers increase on June 1 even though they remain on premium request billing until their term expires. OpenAI's GPT-5.4 jumps from a 1x multiplier to 6x, effectively making each request far more expensive for locked-in annual customers.
Developers pushed back hard in GitHub's community FAQ thread, which drew 70 comments and 105 replies within hours. One user wrote the difference between request-based and token-based billing "could sober up an alcoholic from mere shock," adding: "You will get less, but pay the same price."
Another commenter framed the shift as erasing Copilot's competitive advantage: "This is removing the one real advantage GHCP had over Claude Code et al." Users questioned why they would stay with Copilot if it becomes effectively a wrapper over API consumption without offering direct API controls like max tokens or system prompts.
Business and Enterprise customers get temporary relief: promotional credits worth $30/month for Business and $70/month for Enterprise through June, July, and August. After that promotional window closes sometime after August 2026, Business drops to $19 in credits per seat and Enterprise to $39, exactly matching what each seat costs per month.
Code completions and Next Edit Suggestions remain unlimited across all plans and don't consume AI Credits. But everything else does: chat sessions, agentic workflows, code review (which also burns GitHub Actions minutes), and multi-file editing sessions.
GitHub will release a preview bill in early May so users can see projected costs before the transition hits on June 1.
