Anthropic has pulled Claude Code from its $20/month Pro subscription, and they're calling it a "test." Journalist Ed Zitron spotted the change on Anthropic's pricing page and posted about it on Bluesky on April 21st. The coding assistant remains available on Team plans, but those require a minimum of 5 seats at $20 each. The cheapest path to Claude Code is now $100/month, five times what individual users were paying. Amol Avasare, Anthropic's Head of Growth, told Hacker News commenters the removal is just a test. That's an odd claim given the pricing page has been updated and Pro users are already seeing the feature gone from their accounts. Tests don't usually alter public documentation. The timing makes this worse. Pro users have been reporting severe Claude Opus throttling for weeks. Some claim their 5-hour usage limits disappear after a single question, which would make the service effectively useless for any serious coding work. Pulling the coding tool from individual plans while the core model struggles with capacity looks like either a compute shortage or a deliberate push toward enterprise pricing. Maybe both. Anthropic built its name on coding performance. Removing that from individual plans while Opus throttling makes the remaining product unreliable is a strange gamble. Users in the Hacker News threads are already trying GLM 5.1 and drifting back to OpenAI. If this is a test, it's one Anthropic is losing.