Anthropic Is Testing a Massive Change to Claude Code Pricing—And Pro Users Are Not Happy

Anthropic is quietly running an experiment that's got developers fuming: removing Claude Code from the Pro plan and forcing new users to upgrade to Max ($100/month) to access it. On April 22, the Register caught the company updating its pricing page to show an X (not a checkmark) next to Claude Code for Pro users—a change Anthropic has since tried to walk back by claiming it's only a test affecting ~2% of new signups. But the fact that the documentation was updated site-wide, not just behind a feature flag, tells a different story.

This isn't a sudden decision. Anthropic's growth lead Amol Avasare explained the reasoning: "When we launched Max a year ago, it didn't include Claude Code... Engagement per subscriber is way up. We've made small adjustments along the way, but usage has changed a lot and our current plans weren't built for this." Translation: too many developers are using Claude Code on Pro, and Anthropic can't afford the compute costs at that tier.

Here's the good news: if you have an existing Pro subscription, you're not affected. Avasare explicitly stated that "Existing Pro and Max subscribers aren't affected." This is a test targeting new signups—Anthropic is using A/B testing to see how many developers will accept the $100/month tier instead of dropping to free tier or switching to competitors like OpenAI and Minimax.

But here's the transparency issue: Anthropic updated its live pricing page without warning anyone first. The Claude Code documentation page still lists the feature for Pro. The Claude Code product page still shows Pro access. In some corners of Anthropic's site, Pro still "includes" Claude Code. This scattered approach left developers confused, angry, and forced Avasare to do damage control on X (formerly Twitter) instead of announcing the test properly.

At its core, this is about compute economics. Claude Code is expensive to run. When Opus 4.7 shipped in April, usage exploded—developers started using Claude Code for hours-long agentic tasks, which Max users are already running. The problem: Pro costs $20/month, and developers are consuming tokens worth 5-10x that in real compute value. Anthropic has already introduced usage caps and off-peak discounts. Yanking Claude Code from Pro is the next lever.

Avasare also mentioned that Anthropic is "exploring different options" for subscription tiers. Translation: more experiments are coming. Caps on context windows. Rate limit adjustments. Maybe even per-feature pricing. The company knows it needs to rebalance the math, but it's doing so without clear communication to the developer community—which is a strategic mistake.

If you're a Pro subscriber today, nothing changes in the near term. Avasare promised that "if we do land on something, if it affects existing subscribers you'll get plenty of notice." Whether that's true remains to be seen—but at least there's a buffer.

If you're thinking about upgrading from Claude.ai free tier, the decision just got harder. You now have three tiers: free (no Claude Code), Pro ($20/month, maybe no Claude Code), and Max ($100/month, definitely Claude Code). New developers will see this matrix and a significant portion will drop to free and use cheaper alternatives (Minimax, Qwen, Kimi) instead—which might cost Anthropic more market share than it saves on compute.

For teams: if your company is using Pro, lock in now before this test goes live more broadly. If you're evaluating Claude for internal use, assume Claude Code requires Max pricing. Build that into your cost modeling.

This story isn't really about pricing. It's about communication. Anthropic could have announced this as: "We're testing Claude Code tier placement. Here's why: compute costs. Here's what we're considering. Here's what won't change." Instead, it buried the change in a quietly updated pricing page and only responded when Ed Zitron called it out on X.

Enterprise customers—the ones who actually matter to Anthropic's bottom line—will notice this. They'll worry that Anthropic doesn't value customer communication. That's a much bigger problem than the pricing change itself.