Anthropic Tried to Raise the Price of Claude Code, Then Backed Off in Hours
Anthropic quietly tried to move its popular Claude Code tool out of the cheap plan and into a 100-dollar-a-month tier. The backlash was instant, and the company reversed the change within hours, calling it a limited test. The episode shows how much power AI users now have, and how sensitive pricing has become.
Anthropic quietly tried to move its popular Claude Code tool out of the cheap plan and into a 100-dollar-a-month tier. The backlash was instant, and the company reversed the change within hours, calling it a limited test affecting about 2 percent of new signups. The episode shows how much power AI users now have, and how sensitive pricing has become.
Claude Code is one of Anthropic's most loved products. It lets developers use Claude to write and edit software directly, and it has built a devoted following on the 20-dollar Pro plan. Silently removing it from that plan and reserving it for the 100-dollar Max tier was a fivefold jump for anyone who relied on it, and developers noticed immediately. Loyalty cuts both ways.
The speed of the reversal is the real story. Within hours of the change spreading online, Anthropic walked it back and clarified it was only a test on a small slice of new users. A company moving that fast is a company watching its community closely and afraid of losing goodwill. In a market where rival models are nearly as good, switching costs are low and trust is fragile.
It reflects the squeeze AI companies are under. Running these models is enormously expensive, and the cheap plans that won users may not cover the cost of heavy use, so providers keep testing where they can charge more. Anthropic raising prices on its most intensive tool is a sign that the economics of generous AI subscriptions are getting harder. Someone has to pay for the compute.
The honest read is that this was a small misstep, not a crisis. It touched few users, was reversed quickly, and testing prices is normal business. But it is a useful signal of two things at once: that AI pricing is heading up as the real costs bite, and that users will push back hard and fast when it does. The direction is clear even if this attempt failed.
So a quiet pricing test turned into a public lesson in who holds leverage in AI. A fivefold jump floated and pulled, a community that reacted in hours, and a hint that subscriptions will get pricier over time. The cost of running AI is colliding with what users will accept. Watch for the next, more careful attempt to raise prices.