Anthropic pulled Claude Code from the feature list of its $20-a-month Pro plan on Tuesday, April 21, then spent the next several hours on X explaining that it wasn't really doing that. Not to everyone, anyway. The company's head of growth, Amol Avasare, called it a test affecting roughly 2% of new prosumer signups, and said existing Pro and Max subscribers were untouched.
Then why could everyone see the new pricing grid?
What actually changed
The Pro column on Anthropic's pricing page swapped Claude Code's checkmark for an X. Support documentation, first flagged by Ed Zitron, went from "Using Claude Code with your Pro or Max plan" to "Using Claude Code with your Max plan." One word deleted, whole bundle rearranged.
Max starts at $100 a month (technically Max 5x; there's also a $200 Max 20x for heavier users). That's a 5x jump for anyone who subscribed to Pro specifically to use Claude Code, which, judging from the reaction on Reddit and Hacker News, was a lot of people. And the rollout was messy. The Claude Code product page and the CLI client still referenced Pro access hours after the pricing grid changed. Anthropic eventually reversed the public-facing documentation.
The 2% test that everyone apparently saw
Avasare's post on X framed the change as an A/B experiment: "For clarity, we're running a small test on ~2% of new prosumer signups. Existing Pro and Max subscribers aren't affected." Fine. But if the experiment is 2% of new signups, why was the entire public pricing page updated, globally, on a weekday afternoon?
Simon Willison, writing on his blog, wasn't buying it: "I don't buy the '~2% of new prosumer signups' thing, since everyone I've talked to is seeing the new pricing grid and the Internet Archive has already snapped a copy." He also pointed out that Claude Cowork stayed on Pro, and Cowork is, functionally, Claude Code with a different logo. Make of that what you will.
A company spokesperson did not respond to The Register's follow-up about the discrepancy. Corey Quinn, on X, captured the enterprise-side worry: unannounced documentation changes like this erode trust and push customers toward tool-agnostic exit strategies. That is not a small problem for an AI vendor selling long-term contracts.
The economics Avasare actually admitted to
Strip the "small test" framing from Avasare's thread and the rest reads, to me, like a fairly candid explanation of why the subscription math stopped working. "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, he wrote, "was designed for heavy chat usage, that's it."
Then came Opus 4. Claude Code got bundled into Max and, in Avasare's words, "took off." Cowork launched. Long-running async agents went from novelty to default workflow. "Engagement per subscriber is way up," he wrote. "We've made small adjustments along the way (weekly caps, tighter limits at peak), but usage has changed a lot and our current plans weren't built for this."
Translation: a Pro subscriber running Claude Code all day, or a Max user with a background Cowork agent grinding through tickets, can easily burn compute worth many multiples of the monthly fee. That math doesn't hold forever. Pro was always going to feel the squeeze first.
The competition piled on
OpenAI's Tibo Sottiaux took the moment to post, on brand: "I don't know what they are doing over there, but Codex will continue to be available both in the FREE and PLUS ($20) plans. We have the compute and efficient models to support it." Sam Altman, getting into it, replied to Avasare's thread with two words: "ok boomer." Whether OpenAI's unit economics are actually better or just better-subsidized by Microsoft is a question for another piece.
What happens next
Avasare closed his thread with a promise: "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 posted this, of course, on X.
My read: broader Pro pricing changes are coming, and the 2% experiment is reconnaissance. New Pro subscribers who want Claude Code bundled should probably lock it in before that window closes, or plan for $100 a month. Existing subscribers have a runway. How long a runway, Anthropic hasn't said.
