GitHub Copilot Switches to Token-Based Billing, Users Face Severe Cost Hikes

GitHub has announced a major shift in Copilot's pricing model, moving from a flat monthly subscription to token-based usage billing effective June 1, 2026. The change has sparked significant backlash from the developer community, with some users reporting cost increases of 10 to 60 times their current monthly bills.

What changed

Previously, GitHub Copilot charged a flat rate per user per month (roughly $10 for individuals and $19–$39 for teams). Under the new system, billing is calculated by the number of tokens consumed during code generation and chat interactions. Early reports from developers on Reddit and X suggest the transition is anything but smooth.

One developer claimed their costs would balloon from roughly $29 per month to nearly $750. Another posted a screenshot showing a jump from around $50 to approximately $3,000. "What a joke," wrote one Reddit user. "This new usage model is just stupidly expensive. I'm adjusting mine by cancelling. At that cost, it is no longer cost-effective or useful in any practical way."

Defenders and detractors

Not everyone is critical. Some Copilot users have pointed out that heavy token consumption is often driven by "vibe coding"—extensive AI-generated code without meaningful review. More disciplined developers who use Copilot for targeted completions rather than whole-file generation may see more modest increases.

Nevertheless, the pricing shift has reinvigorated interest in local AI coding assistants. Self-hosted models running through Ollama or LocalAI offer free inference after the hardware cost and work entirely offline, with no per-token billing at all.

What this means for self-hosted AI

The Copilot pricing change is a powerful tailwind for local coding assistants. Open-source models in the 7B–70B parameter range can handle code completion and generation tasks, and tools like Continue.dev with Ollama provide an entirely local Copilot alternative. For teams already managing their own infrastructure, the cost comparison now heavily favours self-hosting.

Earlier this year we covered how to Run Llama 3 Locally with Ollama, which is the foundation for any self-hosted coding assistant. For teams looking to replace Copilot entirely, see our comparison of Open WebUI vs AnythingLLM as alternative frontends for local models.

The bigger picture

GitHub's move follows a broader industry trend away from flat-rate AI pricing and toward consumption-based models. Anthropic, OpenAI, and other providers have all shifted toward token metering. The self-hosted AI community benefits from this trend: every price increase at a major AI vendor makes local alternatives more attractive.

"Vibe coding" and heavy AI reliance in general may need to be re-evaluated as the cost equation changes. For those who can run models locally, the savings are only growing.

Source

**TechCrunch:** https://techcrunch.com/2026/05/30/what-a-joke-github-copilots-new-token-based-billing-spurs-consternation-among-devs/