"How much does Claude Code cost?" is one of the most searched questions about it, and most answers online are out of date. Plans, limits and models have changed several times. Everything below was checked against Anthropic's official pricing page and Claude Code docs on 8 November 2026. Prices are in US dollars and exclude tax. If you are reading this months later, check claude.com/pricing before you pay.
The short answer
Claude Code is included in every paid Claude plan. It is not included in the Free plan. The cheapest way in is Claude Pro at $20 a month.
If you are a student or just starting with AI coding, Pro is enough to learn the tool properly. Most people should stay on it until they hit limits regularly.
Every plan that includes Claude Code
| Plan | Price | Usage | Extras relevant to coding |
|---|---|---|---|
| Free | $0 | Limited | No Claude Code |
| Pro | $20/month, or $17/month billed annually ($200 up front) | 1x (the baseline) | Claude Code, Opus and Sonnet models, Fable only via usage credits |
| Max | $100/month | 5x Pro, every 5 hours and every week | Fable included up to half your weekly limit, $100/month in API credits, priority access at busy times |
| Ultra | $200/month | 10x Pro every week, no 5-hour limit for verified accounts | Everything in Max, $200/month in API credits |
| Team, Standard seat | $20/seat/month annually, $25 monthly | More than Pro | For teams of 2 to 150 |
| Team, Premium seat | $100/seat/month annually, $125 monthly | 5x a Standard seat | Seat types can be mixed, up to $500/month pooled API credits |
| Enterprise | $20/seat/month annually, 20-seat minimum, plus usage at API rates | Scales with usage | SCIM, audit logs, custom retention |
A few details that trip people up:
- Max and Ultra are monthly only. Anthropic's billing table lists annual billing for Pro and Team, not for Max or Ultra.
- Fable is Anthropic's top model. On Pro you pay for it with usage credits. On Max and Ultra it is included for up to half your weekly limit.
- The API credits on Max and Ultra are for building on the Claude Platform (your own apps that call the API). They are a bonus, not extra Claude Code usage.
How usage limits actually work
There is no fixed message count. Anthropic says how fast you reach a limit depends on the model you use, how long your conversations are and how large your files are.
Two windows apply on most plans:
- A rolling 5-hour window. Most accounts have a limit that resets five hours after a session starts.
- A weekly limit on top, on paid plans.
Ultra removes the 5-hour limit for verified accounts and keeps the weekly one.
The part most people miss: chat and Claude Code share one pool. Anthropic's FAQ says activity across Claude on web, desktop, mobile and Claude Code all draws from the same limit. If you spend the afternoon chatting with long PDFs, you have less left for coding that evening.
When you hit a limit, you have three options: wait for the reset, move to a higher plan, or turn on usage credits on a paid plan and keep working at standard API rates.
Subscription or API?
You can also run Claude Code on pay-as-you-go API credits through an Anthropic Console account. Then you pay per token, not per month.
Current API prices for the models you are likely to use, per million tokens:
| Model | Input | Output |
|---|---|---|
| Claude Opus 5.5 | $4 | $20 |
| Claude Sonnet 5.5 | $2 | $10 |
| Claude Haiku 5.5 (prompts up to 100K tokens) | $0.10 | $0.50 |
| Claude Fable 5.1 | $10 | $50 |
Cache reads are much cheaper than fresh input, which matters because coding sessions re-read the same files often.
How much is that in practice? Anthropic's Claude Code docs say that across enterprise deployments, the average cost is around $13 per developer per active day, and $150 to $250 per developer per month, with 90% of users staying below $30 per active day.
Read that next to the plan prices. A developer coding every working day on the API would often spend more than a Pro or Max subscription. Someone who codes two evenings a week might spend less. The API makes most sense for:
- Automation and scripts (CI jobs, scheduled runs) where a personal subscription does not fit.
- Very occasional use.
- Companies that want usage-based billing and spend controls.
Which plan should you pick?
You are a student or learning to code
Pro. It includes Claude Code and the Opus and Sonnet models. Learn to work efficiently first. The habits in how to use Claude Code without wasting tokens will stretch Pro much further than an upgrade would.
If the monthly price is a stretch on a student budget, take the annual plan only once you are sure you will use it, and keep each session short and focused on one task so the limit lasts.
You code most days for work or freelancing
Start on Pro, move to Max when you hit the 5-hour limit most days, not once. Max's 5x usage and included Fable access are worth it when the limit is costing you hours of work every week.
You run long agent sessions or several in parallel
Ultra. No 5-hour limit for verified accounts is the main reason to pay for it. If you run background agents, parallel sessions or long refactors, the 5-hour window is what stops you, and Ultra removes it.
You are a team
Team with mixed seats. Give Premium seats to the people who live in Claude Code and Standard seats to everyone else. Companies that need SCIM, audit logs and per-user spend limits should look at Enterprise, where you pay a seat price plus usage at API rates.
How to make any plan last longer
Your plan matters less than how you use it. The biggest savings come from habits, not upgrades:
- Run
/usagein a session. On paid plans it breaks your usage down by skill, subagent, plugin and MCP server, and flags patterns like long context or frequent cache misses. /clearbetween unrelated tasks. Every request sends the whole conversation, so a session that has collected three tasks re-reads all three on every turn of the fourth.- Pick the model for the job. The docs note that Sonnet costs less on routine work. Save Fable for the hard problems.
- Keep
CLAUDE.mdshort. It loads into every session, so every line costs a little on every task. - Disconnect MCP servers you are not using. Each one adds context.
- Plan first. A clear plan means fewer wrong turns, and wrong turns are the most expensive tokens you will spend.
The cheapest token is the one you never send because you planned the task first.
We go deeper on this in context engineering.
Claude Code vs the alternatives on price
People comparing prices usually look at Cursor and OpenAI's Codex too. The pricing models are different enough that a straight comparison misleads. Cursor is now sold as part of a subscription shared with Grok and X, and Codex is included in ChatGPT plans. We compare them job by job in Claude Code vs Cursor and Codex vs Claude Code.
Before you pay
Three checks before you subscribe:
- Confirm the price on the official page. Plans and limits change. Anthropic states that prices and plans are subject to change.
- Check what tax your card adds. Listed prices exclude tax.
- Try the free workflow first. You can learn the planning and review habits with any AI tool, then subscribe when you know you will use Claude Code daily.
At our free Saturday workshop we set up Claude Code from scratch, build a small project, and show where the tokens go so you can choose a plan with your own numbers instead of guesses. Join this Saturday's session.