GitHub new AI pricing leaves users shocked, bill jumps from Rs 48,500 to Rs 5.13 lakh

Microsoft has shifted GitHub Copilot to a usage-based pricing model. The change has triggered complaints over rising costs and underscored a wider move towards pay-for-use AI.

In Short

• Microsoft replaces fixed Copilot subscriptions with AI credit system
• Users report exhausting monthly credits within hours of usage
• Some estimate costs rising from Rs 48,500 to Rs 5.13 lakh

Microsoft has officially switched GitHub Copilot, its AI coding assistant, to a usage-based pricing model from June 1, replacing the predictable subscription system developers had grown accustomed to. The change means users are no longer simply paying a fixed monthly fee for access. Instead, their usage now depends on how much AI computing power they consume. And many users are discovering that AI can be far more expensive than they expected.

As the new pricing system takes effect, GitHub Copilot users across social media platforms and forums have begun sharing screenshots and statistics showing how quickly they are burning through their monthly AI credits.

For some users, it reportedly took less than a day to consume a significant portion of their monthly allowance.

Many have responded with sarcasm and frustration.

"Last month I used ~60% of my quota in an entire month. This month I've already used 25% on day one. What an amazing improvement," one Reddit user wrote.

Another user said a relatively small coding task consumed a surprisingly large portion of their credits.

"First day of the new pricing model, I tried one prompt to just refine one of my existing change proposal... About 20-30 mins task, ends up burning 16% of my monthly credits in one shot. Bye."

Some users have gone even further, claiming the new pricing structure could dramatically increase their monthly costs.

"It's too expensive for me. In just one day, I used 4,000/20,000 credits, which means I would end up paying $600 a month," another Reddit user said.

What exactly has changed?

Under the previous system, GitHub Copilot subscribers received a certain number of "requests" and "premium requests" depending on their subscription plan. According to GitHub, that approach treated vastly different AI tasks in a similar way. A quick question and a multi-hour autonomous coding session could effectively cost the user the same amount. As a result, GitHub said it was absorbing much of the rising inference cost behind heavy usage.

Inference cost refers to the computing power consumed when an AI model generates responses.

Under the new system, users receive a monthly pool of AI credits. Those credits are consumed based on the amount of computing power required for each task.

The number of credits used depends on factors such as the volume of input and output tokens and the pricing of the underlying AI model being used. That means costs can vary significantly depending on both the complexity of a request and the AI model selected.

Some users see eye-watering estimates

The shift has led some users to calculate what their historical usage would have cost under the new model.

According to screenshots shared online, one user estimated that usage which previously translated to a bill of $500.35 (around Rs 48,500) would now cost $5,290.92 (around Rs 5,13,219) under the new pricing system.

The estimate quickly spread online, becoming another example cited by users worried about the long-term affordability of AI coding tools.

Why Microsoft changed GitHub Copilot's pricing

For years, AI companies largely absorbed the enormous computing costs required to run their models. Every chatbot response, coding suggestion and AI-generated output requires significant computing power, and that comes at a price.

To fund this AI race, companies have relied on investors, cost-cutting measures and, in some cases, workforce restructuring. But as the need for more computing power grows and funding becomes harder to secure, many AI firms are increasingly asking users to shoulder more of those costs and Microsoft is the latest example of this trend.

Looking for alternatives

Predictably, the pricing change has triggered a wave of complaints and cancellation threats. Some GitHub Copilot users say they are considering other AI coding tools, while others are already experimenting with alternatives such as DeepSeek and OpenCode GO.

At the same time, some users are adapting by switching between different AI models depending on the task, allowing them to manage credit consumption more carefully.

The reaction highlights a broader shift taking place across the AI industry. As companies face growing computing costs, the era of AI firms quietly absorbing those expenses may be coming to an end. Instead, users are increasingly being asked to pay based on the actual amount of AI they consume — and many are only now discovering how expensive that usage can be.

- Ends