Anthropic removes Claude Code from Pro plan pricing page, says only 2% of new signups affected by test

Anthropic removed Claude Code from its Pro subscription plan pricing page on Tuesday, though the company claims the change affects only approximately 2% of new prosumer signups as part of a capacity management test.

The pricing page previously stated the Pro plan "includes Claude Code" as of Monday. By Tuesday, that phrase was removed and replaced with an explicit "X" mark in the feature list. The change was first noted by AI industry observer Ed Zitron.

Inconsistent documentation creates confusion

The update did not reach all company documentation simultaneously. The Claude Code product page continued to state the Pro plan provides access to the code generation tool. The CLI terminal output still displayed "Claude Pro" when accessing Claude Code. When queried, Claude.ai itself insisted the Pro plan includes Claude Code, citing documentation pages that mentioned only the Max plan.

Amol Avasare, Anthropic's head of growth, clarified on X that "we're running a small test on ~2 percent of new prosumer signups. Existing Pro and Max subscribers aren't affected." An Anthropic spokesperson did not respond to questions about why a public pricing page update would target only a subset of users.

Capacity constraints drive usage limits

Avasare explained the rationale: "When we launched Max a year ago, it didn't include Claude Code, Cowork didn't exist, and agents that run for hours weren't a thing. Max was designed for heavy chat usage, that's it."

According to Avasare, Claude Code usage surged after Opus 4 launched, and long-running async agents became standard workflows. "Engagement per subscriber is way up," he said. The company has implemented weekly caps and tighter limits during peak hours, but current subscription plans were not designed for observed usage patterns.

The core issue: Anthropic's subscription plans charge substantially less than the token consumption cost—sometimes by a factor of ten or more, according to The Register's analysis. The company introduced usage limits last month to shift demand to off-peak hours.

Customer communication breakdown

Developers in online forums criticized what many initially interpreted as an unannounced plan change. Some noted the move could push users toward Chinese models like Minimax, Qwen, Kimi, and GLM.

Avasare acknowledged the communication failure: "When we do land on something, if it affects existing subscribers you'll get plenty of notice before anything changes. Will hear it from us, not a screenshot on X or Reddit." He delivered this message in a post on X.

Anthropic faces capacity constraints similar to those affecting GitHub and Google. Claude models have grown increasingly popular in recent months, with projects like OpenClaw enabling developers to create agents that consume significant compute resources.

What this means

This episode reveals the economics of AI subscriptions are breaking under actual usage patterns. When subscription revenue fails to cover compute costs by an order of magnitude, providers must either raise prices, reduce access, or accept unsustainable unit economics. The messy rollout suggests Anthropic lacks a clear strategy for managing this transition—concerning for enterprise customers evaluating the platform's stability. Whether this test becomes permanent depends on how the 2% of affected users respond and whether Anthropic can find a pricing model that works for both heavy users and its bottom line.