Anthropic has pulled Claude Code from its $20 monthly Pro plan for new signups, and the move looks more deliberate than the company admits. Support documentation that previously read 'Using Claude Code with your Pro or Max plan' now says only 'Using Claude Code with your Max plan,' according to reporting by Ed Zitron at Where's Your Ed At. Anthropic representative Amol Avasare called it 'a small test on ~2% of new prosumer signups,' but Zitron notes the documentation changes and website updates suggest something broader. Current Pro subscribers still have access through the web app and command-line interface, though whether that grandfathered access persists is anyone's guess.
The economics behind this are straightforward. AI coding tools burn through compute at a rate that makes flat-rate subscriptions hard to justify. The company recently offered free credits to Pro users to celebrate usage bundles, yet that generosity doesn't extend to unlimited access. Claude Code operates as an autonomous agent that reads files, executes terminal commands, and runs tests, with each action triggering a separate inference pass. A single user request can generate dozens of model calls. Add the fact that coding tasks require feeding entire codebases into the model, and computational cost scales quadratically with input length. Twenty bucks a month doesn't cover that for heavy users.
Anthropic already moved enterprise customers from subscription rate limits to per-million-token billing, as The Information reported. Microsoft is heading the same direction with GitHub Copilot, planning token-based billing and tighter rate limits. Two of the biggest names in AI, both backing away from flat-rate pricing for coding tools. For anyone building workflows around these tools, platform dependency carries real risk. The price of convenience can shift without warning.