Anthropic's Pricing Plans Can't Handle Claude Code Workloads
Key Takeaways
- Anthropic briefly removed Claude Code from Pro subscriptions for new users, reversing after criticism
- Head of Growth Amol Avasare admitted current plans 'weren't built for' compute-intensive tool usage
- The incident reflects an industry-wide compute shortage affecting OpenAI and GitHub too
What Happened
Anthropic quietly removed Claude Code from its $20/month Pro subscription for new customers. The change showed up on pricing pages and support documentation before the company reversed it following public criticism.
Ed Zitron, writing at Where's Your Ed At, documented the process. On both mobile and desktop versions of Anthropic's website, the Pro tier showed an X next to Claude Code. The Max 5x and Max 20x plans kept access. Support documentation changed too. An archived version from April 10 carried the headline "Using Claude Code with your Pro or Max plan." The updated version read "Using Claude Code with your Max plan."
Existing Pro users reported they could still access Claude Code through the command-line interface and web app. New signups saw a different story.
Anthropic's Response
Amol Avasare, Head of Growth at Anthropic, responded after the change surfaced on X. He called it a "small test on ~2% of new prosumer signups" and said existing Pro and Max subscribers weren't affected.
Zitron pointed out a problem with this explanation. Both the public pricing page and general support documentation were altered. Every visitor saw Pro listed without Claude Code access. That's not a 2% test.
Anthropic reversed the changes. Avasare apologized for the confusion.
The Real Admission
More interesting than the test itself: what Avasare said about Anthropic's pricing structure. He acknowledged the company had "made small adjustments along the way (weekly caps, tighter limits at peak)" but that "usage has changed a lot and our current plans weren't built for this."
“The Max plan predated Claude Code and Claude Cowork and was designed for heavy chat usage, that's it.”
— Amol Avasare, Head of Growth at Anthropic
This is a clear signal. Anthropic designed its subscription tiers before Claude Code existed. The coding tool's compute demands don't fit into plans built for chatbot conversations. Something has to change.
The Broader Pattern
Anthropic has already shifted enterprise customers to per-million-token billing. Rate limits were often much cheaper than actual usage. Token-based pricing captures the real compute cost.
If Claude Code were permanently dropped from Pro, new subscribers would face two choices: pay for API tokens directly, or upgrade to more expensive plans. Either way, costs go up for coding-heavy users.
Anthropic's new tokenizer adds more expense on top. The company is clearly trying to align revenue with compute costs.
Industry-Wide Compute Crunch
Anthropic isn't alone. OpenAI and GitHub are also restricting services or pausing new signups to conserve capacity. The AI industry is hitting compute limits faster than infrastructure can scale.
For users, this means one thing: the era of cheap, unlimited AI tool access is ending. Providers are moving toward usage-based pricing that reflects actual compute consumption. Flat-rate subscriptions with generous limits were a growth strategy, not a sustainable business model.
Logicity's Take
Compare other AI coding assistants as Claude pricing evolves
Frequently Asked Questions
Is Claude Code still available on the Pro plan?
Yes. Anthropic reversed the removal after public criticism. Both existing and new Pro subscribers currently have access to Claude Code.
Why did Anthropic try to remove Claude Code from Pro?
Anthropic's Head of Growth admitted the current plans weren't built for compute-intensive tools like Claude Code. The Max plan was designed for heavy chat usage before coding tools existed.
Will Claude Code pricing change in the future?
Likely yes. Anthropic has already shifted enterprise customers to per-token billing and acknowledged current plans don't match usage patterns. Restructured pricing seems probable.
Are other AI providers facing similar compute limits?
Yes. OpenAI and GitHub are also restricting services or pausing new signups to conserve compute capacity. This reflects an industry-wide shortage.
Need Help Implementing This?
Source: The Decoder / Maximilian Schreiner
Huma Shazia
Senior AI & Tech Writer
Related Articles
Browse allZuckerberg's Superintelligence Lab Faces Setback
The first AI model from Zuckerberg's superintelligence lab has failed to impress compared to its rivals, sparking concerns about the lab's direction. We take a closer look at what happened and why it matters.

Muse Spark Launch Propels Meta AI App to Top 5
The recent launch of Muse Spark has significantly boosted the popularity of Meta AI app, pushing it into the top 5. We explore what this means for the AI landscape.

Meta's Muse Spark AI Model Lags Behind ChatGPT and Claude
Meta's Muse Spark AI model still can't outperform ChatGPT and Claude in key areas, despite its advancements. We explore what this means for the AI landscape.

Meta Launches Muse Spark AI To Challenge ChatGPT
Meta launches Muse Spark AI to challenge ChatGPT and Claude, we explore what this means for the AI landscape. Muse Spark AI is a significant development in the AI chatbot space.
Also Read
How to Get Google Maps Monthly Travel Recaps
Google Maps offers monthly travel recaps through its Timeline feature, but the experience has changed since 2024. Here's how to enable notifications and what privacy trade-offs you should know about.

Samsung Engineer Gets 7 Years for Selling Chip Secrets to China
A South Korean court sentenced a former Samsung researcher to seven years in prison for leaking 10nm DRAM manufacturing secrets to Chinese chipmaker CXMT. The engineer received $2 million for technology Samsung spent $1.08 billion developing over five years.

Your Loud PC Might Be a Setup Problem, Not a Hardware One
A tech journalist spent money on new fans only to discover her PC noise issues stemmed from case placement and airflow mistakes. The fix cost nothing. Here's what she learned about diagnosing noise before buying replacements.