Anthropic’s Claude Code Removal Sparks Concerns Over Flat-Rate AI Plans as Agentic Usage Grows

Anthropic’s recent test removing Claude Code from its $20-per-month Pro plan has reignited debate over the sustainability of flat-rate AI subscriptions as agentic AI usage surges. The change, which appeared on the company’s pricing page without prior notice, was later clarified as a small-scale experiment affecting approximately 2% of new prosumer signups, with existing subscribers unaffected.

Users first noticed the alteration on April 21, 2026, when the Claude Code feature disappeared from the Pro plan column on claude.com/pricing, replaced by a red X while remaining available in the $100-per-month Max plan. The accompanying support documentation was simultaneously updated to reflect that Claude Code was now exclusive to Max subscribers, signaling a potential shift in feature availability.

Anthropic Head of Growth Amol Avasare addressed the confusion on social media, explaining that the modification was a limited test and not a policy change for current users. He contextualized the decision by noting that when the Max plan launched a year ago, Claude Code did not exist, Cowork was not yet introduced, and long-running AI agents were not common workflows. “Max was designed for heavy chat usage, that’s it,” Avasare stated, adding that engagement per subscriber has since increased significantly.

Avasare emphasized that while Anthropic has implemented incremental adjustments like weekly caps and tighter peak-time limits, the fundamental usage patterns have evolved beyond what the original flat-rate plans were designed to accommodate. “Usage has changed a lot and our current plans weren’t built for this,” he said, acknowledging the growing computational demands of agentic tools like Claude Code and Claude Cowork.

The incident drew parallels to recent moves by GitHub, which suspended signups for its flat-rate Copilot plans and removed Claude Opus access from its $10-per-month Pro tier. These concurrent developments suggest a broader industry reckoning with the mismatch between fixed pricing models and escalating resource consumption driven by advanced AI agents.

Should Anthropic proceed with permanently removing Claude Code from the Pro plan, it would risk alienating subscribers who opted for annual commitments under the expectation of full feature access. Industry observers suggest that any such change should include provisions for existing users to retain functionality or receive refunds to maintain trust.

The vacuum created by restricting access to agentic tools on lower-tier plans could present an opening for competitors. Specifically, OpenAI’s Codex—positioned as a direct alternative to Claude Code—may benefit from user dissatisfaction, particularly if prompted by public remarks from executives like Sam Altman, who responded to Avasare’s comments with an “ok boomer” on social media, hinting at willingness to attract displaced Claude users.

As of the time of reporting, Anthropic had reversed the visible changes to its pricing page and documentation, restoring Claude Code to the Pro plan listing. The company has not announced any immediate plans to alter feature availability for existing Pro or Max subscribers, though the episode underscores ongoing tensions in monetizing AI agents under legacy subscription frameworks.