Claude Code Quietly Removed from Anthropic Pro Plan

On April 22, 2026, Anthropic quietly updated its Claude Pro documentation to remove Claude Code from the listed features — no press release, no in-app notification, no email to subscribers. For the roughly 2% of Pro users caught in the test group, the AI automation coding tool they relied on for daily work simply stopped appearing in their account. That matters because Claude Code has been the single most compelling reason developers chose Claude Pro over the free tier — and losing it silently erodes subscriber trust in ways a formal pricing change never would.

Claude Code: The AI Coding Feature That Vanished Without a Word

Claude Code is Anthropic's built-in coding environment — a workspace inside the Claude interface where you can write, execute, and iterate on code without switching to a separate terminal or IDE (integrated development environment, the app developers use to write software, like VS Code or PyCharm). Unlike generic AI chat, Claude Code maintains persistent context (the ability to remember earlier parts of a long coding session) and can run scripts in real time, not just suggest code and leave execution to you.

Subscribers who lost access received no downgrade notice and no refund adjustment. The removal was detectable only through changes to Anthropic's public-facing help documentation — the pages that describe what each subscription includes. Because the documentation update rolled out globally rather than just to the test cohort (the specific group of users selected for the experiment), every Claude user could see the change, not just the 2% who were actually affected.

Anthropic has since acknowledged the change is a limited A/B test (a controlled experiment where two different versions of a product are shown to separate user groups to measure their reactions). No formal statement accompanied the test launch. No timeline has been published for when the test ends or what decision it will inform.

$800 Billion in Funding — and Claude Code Quietly Removed

The timing is difficult to overlook. April 22, 2026 is the same date Anthropic finalized its $25 billion investment from Amazon, pushing the company's valuation to approximately $800 billion — making it one of the most valuable private AI companies in history. A company that just secured generational funding, quietly running an experiment to see what happens when a flagship Pro feature disappears, signals something well beyond routine product maintenance.

Three scenarios are credible given what's publicly visible:

• Pricing tier restructure — Claude Code may be moving to its own paid tier (modeled after how GitHub Copilot is sold separately from GitHub's base subscription). The test measures how many Pro subscribers stay, leave, or upgrade before any price announcement is made public.
• Compute capacity management — Running code in real time is resource-intensive (it requires dedicated server infrastructure that goes far beyond standard language model inference). GitHub Copilot paused all new individual plan sign-ups in April 2026 due to capacity constraints, suggesting this pressure is industry-wide. Anthropic may be evaluating whether Claude Code's compute footprint is financially sustainable at Pro-tier pricing.
• Portfolio rebalancing after Claude Design — In the exact same week, Anthropic launched Claude Design, a new tool for creating UI prototypes (visual mockups of apps and websites) without Figma or traditional design software. Product line refreshes often involve reassigning features between tiers to create cleaner upgrade paths for users.

Two Percent Test, One Hundred Percent Documentation Confusion

The disconnect between the test's scope and its documentation rollout is where the communication strategy breaks down entirely. A properly scoped A/B test would apply both the feature change and the supporting documentation only to the test cohort. Here, the documentation — the most visible signal a subscriber has about what they're paying for — changed for everyone, while the actual feature removal only hit 2% of accounts.

The practical effects of this gap were immediate:

• Pro subscribers outside the 2% test group saw help pages suggesting Claude Code was no longer a Pro feature — before they could know they still had full access
• Developers relying on Claude Code for production workflows (active projects deployed in live software environments) may have already begun switching to alternatives based on documentation alone
• No official channel existed to help affected subscribers distinguish a product change from a billing error or a temporary technical outage

Tech reporter Thomas Claburn at The Register was first to document the change on April 22, publishing without an Anthropic statement available. The absence of any proactive communication from a company that just closed a $25 billion investment round is unusual by Anthropic's own historical standards of transparency around pricing shifts.

AI Automation Market Pressure Driving Anthropic's Decision

Anthropic is not operating in isolation. The AI coding tool market shifted significantly in the weeks surrounding this test:

• GitHub Copilot halted new individual plan subscriptions in April 2026 due to server capacity issues — signaling that demand for AI coding tools is outrunning the infrastructure supporting them at consumer-grade prices
• OpenAI Codex Desktop launched as a free alternative to paid AI coding environments, applying direct pricing pressure on any coding feature locked behind a monthly subscription paywall
• Anthropic's own Opus 4.7 launched at $5 per million tokens (a unit measuring how much text an AI processes — roughly 750 words per token batch) but runs approximately 40% more expensive in practice than its predecessor, suggesting the company is already managing cost-per-use pressure across its product line

The $25 billion Amazon investment introduces another strategic dimension. Large enterprise deals (contracts with corporations using AI at scale, rather than individual subscribers) often carry expectations about which capabilities stay at consumer pricing versus dedicated enterprise pricing tiers. Claude Code's removal from the $20/month Pro plan could be the first visible step in repositioning it as an enterprise-exclusive or API-only capability — a move that would meaningfully boost per-seat revenue without touching the headline Pro price.

Check Your Claude Pro Dashboard Right Now

If you're a Claude Pro subscriber, the fastest way to confirm your status is to log into claude.ai and look for Claude Code in your available tools panel. If it's missing and you were relying on it for active projects, three immediate actions are available:

1. Contact Anthropic support directly, reference the April 22 A/B test, and request reinstatement or a clear status update for your account
2. Explore alternative AI-powered coding setups and AI automation guides — including local models that run entirely on your own machine with no subscription — as a bridge while Anthropic finalizes its pricing structure
3. Watch Anthropic's official channels for a formal announcement; the combination of the Amazon investment close, the Claude Design launch, and this test all point toward a product pricing update coming within weeks

The broader pattern here extends well beyond this single test. AI companies are actively reshaping which features belong at which price points as their enterprise ambitions and valuations grow. What's bundled into a $20/month plan today may require a separate subscription — or a significantly higher tier — before 2026 ends. If your team is building AI automation workflows that depend on any AI coding tool, auditing pricing terms every quarter is no longer optional. Start with your Claude account right now.

Related Content — Get Started | Guides | More News