GitHub Copilot shifts to token-based billing

Microsoft is transitioning its AI-powered coding assistant, GitHub Copilot, from a flat subscription rate to a token-usage billing system. The changes will take place June 1, meaning users will be billed based on the volume of tokens they consume during development rather than paying a predictable flat fee. The new billing system has the potential to bill users at a significantly higher rate, raising concerns among developers who rely on the tool for daily workflows.

Some users are experiencing what appears to be drastic cost increases, taking to platforms like Reddit and X to share their financial whiplash. The reported projections show steep climbs from previous flat rates:

• One user reported that their monthly cost is projected to jump from around $29 to nearly $750.
• Another user shared a screenshot showing a projected monthly cost increase from around $50 to some $3,000.

In response to these projections, one Redditor dismissed the change as a joke. According to one Redditor, “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.” Another user expressed surprise at the scale of the increase, calling the new pricing model ridiculous.

The developer community is divided on what is driving these pricing spikes. Some users argue that the extreme bills are the result of inefficient or non-technical use of AI, a practice referred to as vibe-coding. One user pointed out the stark contrast between developers who work all day without exceeding limits and those facing massive bills, questioning whether workload complexity alone could explain the difference. The same user suggested that such extreme costs only occur when users engage in vibe-coding with a high volume of bloated iterations, adding that the service remains affordable for small teams when treated as a standard tool. Other observers have questioned how much money the assistant was losing under the previous flat-rate subscription model.

Conversely, other developers argue that Microsoft is at fault for encouraging high-token usage patterns. One user argued that Microsoft provided this billing method and made it increasingly easy to burn through massive numbers of tokens on single premium requests. These requests could run for hours or days while spawning dozens or hundreds of sub-agents—which are secondary AI agents spawned by a primary request—leading to the massive token consumption now being billed to users.

Why it matters

The shift from flat-rate to usage-based pricing signals a broader trend in AI monetization, forcing developers to balance the utility of AI coding assistants against potentially unpredictable and significantly higher operational costs.