GitHub halts Copilot Pro signups as agentic AI workloads overwhelm infrastructure
GitHub has paused new subscriptions for Copilot Pro, Pro+, and Student plans due to compute capacity constraints. The company cites agentic workflows as consuming significantly more resources than its original pricing structure anticipated, forcing tighter usage limits and a shift away from flat-rate billing.
GitHub halts Copilot Pro signups as agentic AI workloads overwhelm infrastructure
Microsoft's GitHub has stopped accepting new Copilot individual subscriptions as agentic AI workloads consume far more compute resources than the company's pricing model can support.
GitHub paused signups for Copilot Pro, Pro+, and Student plans on April 20, 2026, according to Joe Binder, VP of product. The company said it needs to serve existing customers more effectively as long-running, parallelized agentic sessions "regularly consume far more resources than the original plan structure was built to support."
Capacity crisis spreads across AI industry
The move follows similar capacity constraints across the AI infrastructure landscape. Anthropic adjusted usage limits in February to shift consumption away from peak hours following the OpenClaw surge. Google implemented similar policies for Gemini CLI and Gemini Code Assist. OpenAI enacted usage balancing measures earlier in April. AWS reportedly lost business to Google Cloud in 2025 due to inability to meet AI demand.
GitHub suspended Copilot Pro free trials last week due to abuse. The free tier remains available.
Usage limits tightened, premium models removed
GitHub is tightening both session and weekly usage limits tied to token consumption. Session limits cap usage during peak times, forcing users to wait until the usage window resets. Weekly limits control "parallelized, long-trajectory requests that often run for extended periods of time and result in prohibitively high costs," according to Binder.
The company is removing Anthropic's Opus 4.5 and 4.6 models from Pro+ subscriptions. Opus 4.7, launched last week, will be available to Pro+, Teams, and Enterprise customers with a 7.5× premium request multiplier through April 30. The discontinued Opus 4.6 carried a 3× premium. Opus 4.7 is 20-40% more expensive than predecessors but performs better in certain scenarios, according to GitHub.
Shift to token-based billing
The changes reflect a transition from flat-rate per-request billing to token-based pricing. GitHub currently bills per interaction regardless of token consumption, which can cost Microsoft more than it charges when requests trigger extended reasoning chains.
Pro and Pro+ subscribers have until May 20 to adjust to the new limits. The company did not provide a timeline for reopening signups.
What this means
GitHub's signup pause exposes the fundamental mismatch between agentic AI economics and current pricing models. Flat-rate billing breaks when autonomous agents generate unpredictable, long-running workloads that can cost providers orders of magnitude more than they charge. The industry-wide capacity crunch—from Anthropic to Google to AWS—indicates this isn't a temporary blip but a structural problem as AI systems shift from simple completions to complex, multi-step reasoning. The move toward token-based pricing is inevitable, but providers face a tough sell convincing users to accept higher, less predictable costs.
Related Articles
Taste-Bench: Best AI model scores 59.7% on long-term agent decisions, where random guessing yields 25%
Researchers from Microsoft and City University of Hong Kong released Taste-Bench, a 502-question benchmark that tests whether models can pick the better path at decision forks in long agent runs. The top model, GPT-5.6 Sol, scored 59.7% against a 25% guessing baseline. The authors also show a distilled Qwen3.6-27B student can lift SWE-bench Pro success from 14.6% to 33.7%.
Hugging Face Transformers v5.20.1 adds Reflection AI's Beam model and fixes loading on older torch
Hugging Face's Transformers v5.20.1 adds Beam, a model from Reflection AI that the release notes describe as token-efficient, frontier-scale, and aimed at coding and agentic work. The patch also fixes a regression that broke model loading on older torch versions. Beam's specifications and pricing were not disclosed in the release notes.
Microsoft's Probabilistic Bias Correction improves ECMWF forecasts for 98% of precipitation targets, runs on CPUs
Microsoft Research has published Probabilistic Bias Correction (PBC) in Nature and released it as open-source software. The researchers report it improved a leading ECMWF physics-based model for 98% of precipitation targets and 92% of temperature targets at 2–6 week lead times, and that it runs on CPUs.
Microsoft releases Decision-1, a Qwen3.5-9B-based model for classification and routing, at $0.042 per 1M input tokens
Microsoft has released Decision-1, a decision model built on Qwen3.5-9B for classification, evaluation, and routing. Microsoft claims 83.5% accuracy across 36 benchmarks and 85 ms latency. Input tokens cost $0.042 per million, and output tokens are free.
Comments
Loading...