“Although I was very cautious on the first day, testing it with a limited number of uses, it still consumed 840 credits,” said one user. wrote Try Claude Sonnet 4.6 through Copilot today. “I haven’t even done any really complex work yet,” another user complained later reported usage accounts for 21 percent of your monthly Pro Copilot subscription credit allocation in a single day. “I have a feeling I’ll be going somewhere else soon.”
Using your organization’s 8,000 monthly AI credits in a single day is…probably not sustainable.
Amid the price change, many GitHub Copilot users are predictably publicly threatening to cancel their subscriptions or pursue other AI coding options. But others say they have been able to adapt to the new world of usage-based pricing. Henri Kinnunen coder writes that they only burned 161 credits in a “productive day” of using GPT 5.3-Codex through Copilot, thanks to the fact that they limited themselves to “very focused and deliberate changes with AI.” At Bluesky, coder Neil Hewitt wisely observed that continuing a three-day chat session in Copilot probably isn’t so smart now, since it means sending “the entire chat history as context every time… hey, input tokens use credits… it’s not rocket science.”
While some Copilot users are jumping ship for other services with more generous usage limits, that kind of subsidized customer acquisition could soon give way to Copilot-style usage-based pricing across the industry. If that happens, LLMs that are more efficient with their tokens may win the economic battle; on Reddit, a user is already arguing how they have integrated Deepseek into their GitHub VSCode environment at a cost of only “about 7 cents for 15 million tokens.” While one could say “you get what you pay for,” some AI users are now contemplating a world in which they, too, will have to pay for what they get.
