If you’ve been using Claude Code on Anthropic’s $20/month Pro plan, that door just closed. Anthropic quietly updated its pricing page to remove Claude Code from the Pro and Max tiers entirely. It’s now only available on Team ($20 to $100/seat/month) and Enterprise plans.
This matters if you run a service business. A lot of solo operators and small teams had been leaning on Claude Code as their AI coding assistant for internal tools, automations, and client work. That access is gone unless you upgrade.
What happened
- Claude Code has been removed from the Pro ($20/month) and Max (from $100/month) individual plans. The feature comparison table on Anthropic’s pricing page now shows Claude Code only under Team and Enterprise.
- Team plan pricing starts at $20/seat/month (standard) or $100/seat/month (premium). Premium seats include Claude Code and Claude Cowork. You need a minimum of 5 seats.
- Enterprise plans charge $20/seat plus usage at API rates. Claude Code is included, along with SCIM, audit logs, HIPAA readiness, and other compliance features.
- The Pro plan still includes Claude Cowork, web search, Research, desktop extensions, and projects. But not Claude Code.
- Free plan users were never included, so nothing changes there.
The numbers
- Pro plan: $17/month billed annually ($200 upfront), or $20/month billed monthly. No Claude Code.
- Max plan: From $100/month for 5x or 20x more usage than Pro. Still no Claude Code.
- Team standard seat: $20/seat/month (annual) or $25/month (monthly). Claude Code not included at this tier.
- Team premium seat: $100/seat/month (annual) or $125/month (monthly). Claude Code included.
- Enterprise: $20/seat plus API usage rates. Claude Code included.
- API rates for reference: Opus 4.7 runs $5 input / $25 output per million tokens. Sonnet 4.6 runs $3 input / $15 output. Haiku 4.5 is $1 input / $5 output.
5 things service business operators should do right now
- Audit your Claude Code usage. If your team has been using Claude Code on Pro accounts, figure out how many people actually depend on it and for what tasks. You can’t make a smart decision without knowing the scope.
- Do the math on Team premium seats. At $100/seat/month for premium (which includes Claude Code), a 5-person team is looking at $6,000/year. Compare that to what you’d spend on API access or alternative tools for the same coding workflows.
- Evaluate the Enterprise plan’s usage-based model. If your Claude Code usage is bursty (heavy some weeks, light others), the Enterprise plan’s $20/seat plus API rates might actually be cheaper than flat-rate premium seats. Model it out.
- Test alternatives now, not later. OpenAI’s coding capabilities, GitHub Copilot, Cursor, and open-source options like Continue.dev all compete in this space. Don’t wait until your workflows break. Run a two-week trial on a real project.
- Redesign workflows around what’s still free or cheap. Claude Cowork is still on the Pro plan. So is Research, web search, and code execution in chat. Some tasks you were doing in Claude Code might work fine through those remaining features with a little restructuring.
The hot take
This is Anthropic telling you exactly where they think the money is. Coding agents are too valuable to give away at $20/month. Every AI company is going to do this. The pattern is predictable: launch a powerful feature broadly to build adoption, then gate it behind higher tiers once people depend on it. If your business workflows rely on any single AI vendor’s bundled features, you’re building on rented land. The rent just went up.
The Agency OS play
This week, sit down and list every workflow in your business that touches Claude Code. Be specific. Is it generating boilerplate for client projects? Writing internal scripts? Debugging production code? For each one, write down two things: how often it runs, and what it would cost you in time if the tool disappeared tomorrow.
Then build a simple spreadsheet comparing three paths. Path one: upgrade to Team premium seats at $100/seat/month. Path two: move coding tasks to the API directly (where you pay per token and can use whichever model fits the job). Path three: switch those specific tasks to a competing tool. For most small service businesses, the answer will be a mix. Keep Claude Pro for the chat and research features your team loves. Route the heavy coding work to the API or a dedicated coding tool where you control costs.
If you manage a team of five or more, this is also a good forcing function to standardize your AI tooling. Half your people are probably on individual Pro accounts billed to personal cards. Consolidate. Whether you go Team or Enterprise or multi-vendor, centralizing your AI spend gives you visibility into what you’re actually paying and what you’re actually getting. Do it before the next price change catches you off guard.