GitHub Copilot shakes up individual plans: new sign-ups paused

GitHub has announced a set of changes to its Copilot individual plans, aimed at protecting service quality for existing subscribers. The move comes as agentic workflows place unprecedented demands on the platform's infrastructure, with long-running, parallelized sessions consuming far more compute resources than the original plan structure anticipated.

Three key changes take effect immediately:

  1. New sign-ups for GitHub Copilot Pro, Pro+, and Student plans are paused, allowing GitHub to prioritize serving its current customer base.
  2. Usage limits for individual plans are tightened, with Pro+ offering more than 5X the limits of Pro. Usage limits are now displayed in VS Code and Copilot CLI to help users avoid hitting them.
  3. Opus models are removed from Pro plans. Opus 4.7 remains available on Pro+, while Opus 4.5 and Opus 4.6 will be phased out of Pro+ as previously announced in the changelog.

Users affected by these changes can cancel their Pro or Pro+ subscription and receive a refund via their Billing settings before May 20. Monthly plans get the full current month's fee refunded; annual plans receive a prorated refund for remaining time.

How usage limits work

Copilot employs two distinct usage limits: session limits and weekly (7-day) limits. Both depend on token consumption and the model's multiplier.

Session limits prevent service overload during peak demand and are calibrated so most users remain unaffected. Users hitting a session limit must wait until the usage window resets.

Weekly limits cap total token consumption over seven days, addressing the high costs associated with parallelized, long-trajectory requests. These limits also target minimal impact on typical users. If you exhaust your weekly limit but still have premium requests remaining, Copilot automatically switches to Auto model selection until the weekly period resets.

Usage limits operate independently of premium request entitlements. Premium requests govern model access and request counts, while usage limits are token-based guardrails within set time windows. It's possible to have both premium requests remaining and exceed a usage limit simultaneously.

Tools for transparency and managing limits

VS Code and Copilot CLI now display available usage as you approach limits, giving you advance notice before disruption. For users nearing their cap, GitHub suggests several strategies:

  • Choose models with smaller multipliers for simpler tasks. Larger multipliers accelerate limit consumption.
  • Upgrade from Pro to Pro+ to multiply your limit by over 5X.
  • Leverage plan mode in VS Code or Copilot CLI to improve task efficiency and success rates.
  • Reduce parallel workflows. Tools like /fleet consume significantly more tokens and should be used sparingly when close to limits.

Justification for the changes

Usage intensity has risen across the board as users embrace agents and subagents for complex coding tasks. These workflows deliver real value but strain infrastructure and pricing models—in some cases, a handful of requests can incur costs exceeding an entire plan price. GitHub acknowledges these are issues to solve internally while ensuring existing users receive a predictable experience. The current actions create headroom to develop a more sustainable long-term solution.