GitHub Pauses Copilot Signups, Tightens Usage Limits
GitHub is pausing new signups for individual Copilot plans and tightening usage limits to prioritize service reliability for existing customers. Paid tiers affected are Copilot Pro and Copilot Pro+ plus the Student plan; Copilot Free remains open. Model access is being narrowed: Opus models are being restricted, Opus variants are no longer available on Pro, Opus 4.7 will be restricted to Pro+, and Opus 4.5 and Opus 4.6 will be removed from Pro+. GitHub will surface usage limits in editors such as VS Code and the Copilot CLI, and Pro+ offers more than 5X the limits of Pro. Existing Pro and Pro+ subscribers who find the changes unacceptable can cancel and request a prorated refund through May 20, 2026. The changes are a response to rising compute pressure from agentic, long-running sessions that drove recent incidents and degraded service.
What happened
GitHub announced on April 20, 2026 that it is pausing new signups for individual Copilot plans and tightening usage limits to protect service reliability for existing paying customers. The pause affects signups for Copilot Pro, Copilot Pro+, and Student plans; Copilot Free remains available. Model availability is changing: Opus models are no longer available on Copilot Pro; Opus 4.7 remains available on Pro+, and Opus 4.5 and Opus 4.6 will be removed from Pro+.
Technical details
The company cites growth in agentic, long-running, parallelized sessions as the root cause of rising compute demand and recent incidents that degraded performance. Key operational changes include:
- •Pausing new sign-ups for Copilot Pro, Copilot Pro+, and Student plans to prioritize existing customers.
- •Tightening usage quotas; Pro+ plans will offer more than 5X the limits of Pro. Usage warnings will appear in VS Code and the Copilot CLI, and usage progress tracking is coming soon.
- •Removing lower-tier access to older Opus models; Opus 4.7 is retained on Pro+, while Opus 4.5 and Opus 4.6 will be removed from Pro+ over time.
Billing and refund mechanics are explicit: users who find limits unacceptable can cancel Pro or Pro+ and request a prorated refund through May 20, 2026 via Settings → Billing and licensing → Licensing → Manage subscription → Cancel and refund subscription. GitHub warns of temporary authorization holds for usage-based costs and clarifies premium request quotas renew on the first of the month.
Context and significance
This is a product-level throttling and resegmentation response to a classic infrastructure scaling problem: usage patterns evolved faster than quota and pricing models. The shift toward agentic workflows means sessions can consume long-lived compute and parallel inferencing, which is expensive and fragile for shared, flat-rate plans. The changes follow multiple incidents in March that impacted GitHub services and mirror industry moves toward metered usage and quota enforcement seen across developer AI tooling.
For developers, this alters capacity planning. Projects that previously relied on flat monthly plans may now hit hard limits or incur additional charges if they need premium requests beyond allotted quotas. The explicit UI integrations for usage warnings in VS Code and the Copilot CLI are pragmatic: surface limits early to avoid silent failures in developer workflows.
What to watch
Monitor how GitHub implements usage progress tracking and quota enforcement thresholds in editors and CI. Watch pricing and quota signals from competitors and model providers; tokenized or metered models will pressure teams to instrument usage and selectively route heavy workloads to dedicated infrastructure or Pro+ level subscriptions.
Bottom line
This is a defensive operational move that will force developers to treat Copilot usage as a metered resource. Expect increased emphasis on instrumentation, quota-aware design in coding agents, and potential migration paths to team or enterprise plans for heavy agentic workloads.
Scoring Rationale
This is a notable product change that affects many individual developers and workflows, forcing operational changes and capacity planning. It is not a frontier model or regulation shift, but it materially changes access and billing for a widely used developer tool. Freshness adjustment applied.
Practice with real Payments data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all Payments problemsStep-by-step roadmaps from zero to job-ready — curated courses, salary data, and the exact learning order that gets you hired.


