Microsoft to Shift GitHub Copilot to Token-Based Billing, Tighten Rate Limits

Microsoft has announced significant changes to its GitHub Copilot AI coding assistant, including a shift to token-based billing and tighter rate limits. These adjustments come as the company aims to better align user costs with the actual compute expenses of running advanced AI models.

Executive Summary

• Microsoft plans to temporarily suspend new signups for student and paid individual tiers of GitHub Copilot.
• The company will transition from request-based to token-based billing, reflecting the true cost of usage.
• Rate limits for both individual and business accounts will be tightened, and access to certain models will be restricted for users with cheaper subscriptions.

Why it Matters

The shift to token-based billing is a strategic move by Microsoft to manage the growing costs associated with running GitHub Copilot. According to internal documents reviewed by Where’s Your Ed At, the weekly cost of operating the service has nearly doubled since January 2026. This trend mirrors the broader challenges faced by AI companies, which have been subsidizing compute costs to keep user fees low.

Token-based billing will ensure that users pay for the actual computational resources they consume, rather than a fixed number of requests. For instance, more complex models like Claude Opus 4.7 currently cost $5 per million input tokens and $25 per million output tokens. This pricing model is designed to be fairer and more sustainable, as it directly correlates with the compute usage.

Key Risks

1. User Dissatisfaction: The transition to token-based billing and tighter rate limits may lead to user dissatisfaction, particularly among those who have grown accustomed to the current request-based system. Users might perceive this change as a cost increase, potentially leading to churn.
2. Market Competition: Other AI coding assistants could capitalize on Microsoft's changes by offering more flexible or cost-effective plans. Competitors like Anthropic and OpenAI, which are also facing similar compute costs, may find ways to attract users who are dissatisfied with GitHub Copilot's new pricing model.
3. Implementation Challenges: The transition to token-based billing requires significant backend changes and user education. Ensuring a smooth rollout without disrupting existing workflows will be critical.

The Opportunity

1. Sustainable Growth: By aligning costs more closely with usage, Microsoft can ensure the long-term sustainability of GitHub Copilot. This approach could lead to better resource allocation and improved service quality.
2. Premium Services: The company has the opportunity to introduce premium tiers that offer enhanced features and higher token allowances. These premium plans could attract users who are willing to pay more for advanced capabilities.
3. Innovation and Development: With a more stable financial foundation, Microsoft can invest in further developing GitHub Copilot. This includes integrating new models, improving performance, and adding innovative features that enhance the user experience.

Current Billing Model

At present, GitHub Copilot users are allocated a certain number of "requests" based on their subscription tier:

• Pro ($10/month) accounts receive 300 requests per month.
• Pro+ ($39/month) accounts receive 1500 requests per month.

More expensive models use more requests, while cheaper ones use fewer. For example, a complex model might consume multiple requests for a single interaction, whereas a simpler model might require only one request.

Transition to Token-Based Billing

The shift to token-based billing will see users charged based on the number of tokens their prompts and outputs generate. This change is expected to be more transparent and fair, as it directly reflects the computational resources consumed. However, the exact timing for this transition has not been specified by Microsoft.

Source