Microsoft to Shift GitHub Copilot to Token-Based Pricing
Token billing expected to start on June 1
Microsoft is preparing a major pricing change for GitHub Copilot, moving away from fixed monthly limits as demand and infrastructure costs continue to surge, according to Where’s Your Ed At.
Token-based billing replaces request limits
The current $20–$30 monthly pricing model is no longer sustainable at scale. According to reports, Microsoft plans to introduce token-based billing starting June 1.
Instead of capped requests, users will pay based on actual usage. Pricing is expected to land around $2.50 per million input tokens and $15 per million output tokens. This aligns Copilot more closely with how modern AI services charge for compute.
Subscriptions will remain in place, but they will now include token credits rather than fixed usage limits.
New structure for individual and enterprise users
Under the new system, monthly plans will act as a base layer that provides credits. Once those credits are used, additional usage will incur extra costs.
For enterprise customers, Microsoft is expected to introduce pooled credits across teams. This allows organizations to distribute usage more flexibly instead of assigning strict per-user limits.
Reported pricing suggests Copilot Business could offer around $30 in credits for $19 per month, while Copilot Enterprise may provide roughly $70 in credits for $39 per month.
Why Microsoft is making the change
The shift reflects a broader trend in AI pricing. As tools like GitHub Copilot evolve from simple code autocomplete into full AI agents, compute requirements increase significantly.
Token-based billing gives Microsoft a way to better align revenue with actual infrastructure usage. Heavy users will pay more, while lighter users may end up paying less.
The trade-off comes in predictability, as monthly costs will now vary depending on usage patterns.
Bigger push toward AI infrastructure
This move fits into Microsoft’s wider AI strategy, which includes increasing investment in infrastructure and services. The company has already limited access to GitHub Copilot in some cases due to high demand.
At the same time, Microsoft is introducing new AI capabilities, such as hosted agents in its Foundry Agent Service, and implementing workforce changes, including voluntary buyouts in the U.S.
Together, these moves highlight a clear direction: scaling AI products requires both new pricing models and significant resource reallocation.
Via Neowin
Read our disclosure page to find out how can you help Windows Report sustain the editorial team. Read more
User forum
0 messages