Claude Code Pricing: How Much It Really Costs in 2026

Claude Code pricing starts at $20 a month. It comes with every paid Claude plan: Pro is $20 ($17 billed yearly), Max is $100, Ultra is $200 and Team seats start at $25. The Free plan doesn't include it. On the API you pay per token, which Anthropic says averages about $13 per developer per active day.

This guide covers what Claude Code costs on each plan, how API billing works, what usage really looks like, how to check yours and how to spend less. Want your own number? Use the free Claude Code cost calculator.

Prices come from Claude's pricing page and API pricing docs. Usage figures come from Claude Code's costs docs. All were checked November 5, 2026. The pricing page can show local currency, so you may see different numbers outside the US.

Two ways to pay

A Claude subscription

A fixed monthly price with usage limits

• Pro, Max, Team or Enterprise
• Usage resets on a rolling window
• Predictable bill, usage credits optional

vs

The API (Claude Console)

Pay for the tokens you use

• No plan limits, billed per token
• Workspace spend limits available
• Required for agents you build for clients

Claude Code pricing by plan

There's no separate Claude Code subscription. You use Claude Code through a Claude subscription, and each tier raises your limits. Claude Pro is the entry tier. Claude Max and Ultra are for heavy daily coding.

From Anthropic's pricing page, in US dollars before tax, November 2026:

Plan	Price	Claude Code
Free	$0	Not included
Pro	$20/month billed monthly, or $17/month billed yearly ($200 up front)	Included
Max	$100/month	Included, 5x Pro usage
Ultra	$200/month	Included, 10x Pro usage each week
Team, Standard seat	$25/month billed monthly, or $20 billed yearly	Included
Team, Premium seat	$125/month billed monthly, or $100 billed yearly	Included, 5x a Standard seat
Enterprise	$20 a seat a month billed yearly, plus usage at API rates	Included

Only Pro and Team have yearly pricing. Enterprise has a 20 seat minimum. Max also includes $100 a month in API credits and Ultra includes $200, per the pricing page.

Claude Code shares your plan's limits with Claude chat, so your work in both draws from one pool. Most limits reset on a rolling five-hour window, and paid plans add a weekly limit, which works as a weekly cap on top of the five-hour window. Ultra has no five-hour limit for verified accounts. When you hit a limit, it tells you when it resets. You can turn on usage credits to keep working at standard API rates, with a spend limit you set.

API billing: what usage really costs

With an API key from the Claude Console, you're billed by the tokens it uses. How much depends on the model, the size of your codebase and how you work.

Current rates for the main models, in US dollars per million tokens:

Model	Input	Output	Cache read
Claude Sonnet 5.5	$2	$10	$0.10
Claude Opus 5.5	$4	$20	$0.20
Claude Fable 5.1	$10	$50	$0.25
Claude Haiku 5.5 (prompts up to 100K tokens)	$0.10	$0.50	$0.01

Anthropic calls Claude Opus 5.5 its daily driver for agentic coding. Claude Sonnet 5.5 costs half as much per token.

Cache reads matter here. Your conversation goes out with each request, and the sample session in Anthropic's docs is mostly cache reads.

Anthropic's docs give these figures from enterprise deployments:

• About $13 per developer per active day
• $150 to $250 per developer per month
• Under $30 per active day for 90% of users

Those are averages across companies, not a quote for you. Put your own usage into the cost calculator, then run a small pilot and measure before you budget for a team.

What makes usage go up

From the docs, the big drivers are:

• Long sessions. The conversation goes out with each request, so a session open all day costs more per question than a fresh one.
• The model. Opus costs more than Sonnet. The docs suggest Sonnet for most coding and Opus for complex architecture or multi-step reasoning.
• Breaks longer than the cache lifetime. Your first message after a long break reprocesses the full context.
• Subagents, agent teams and scheduled tasks, which each send their own requests.
• Vague prompts. "Improve this codebase" makes Claude read widely. A specific request reads less.

How to check your usage

Type /usage in Claude Code. On Pro, Max, Team and Enterprise plans it shows your plan usage bars and a breakdown of what's using them: skills, subagents, MCP servers and scheduled loops. With an API key, it shows the session's token use and an estimated dollar cost. The authoritative figure for API billing is the Usage page in the Claude Console.

In scripts, claude -p with --output-format json includes a total_cost_usd estimate for each run (headless mode).

How to spend less

The docs' advice, in order of impact for most people:

1. /clear between unrelated tasks. Stale context costs tokens on every message.
2. Match the model to the job. Use /model to switch. Keep Opus for hard problems.
3. Write specific prompts. Name the file and the change.
4. Plan first on bigger changes. Plan mode avoids expensive rework when the first approach is wrong.
5. Keep CLAUDE.md short (under about 200 lines) and move specialist instructions into skills, which load only when needed.
6. Send noisy work to subagents, so long logs and search results stay out of your main context.
7. Disable MCP servers you aren't using with /mcp.

Claude Code pricing vs Cursor, GitHub Copilot and Codex

Every coding agent here has a monthly plan with usage limits, so compare what you get for the money, not only the price. These are individual plans in US dollars, checked November 6, 2026.

Tool	Paid plans	How usage works
Claude Code	$20, $100, $200	Shared Claude limits, reset every five hours and weekly
Cursor	$8, $30, $100, $200	Usage pool; the $8 Plus plan is Grok models only
GitHub Copilot	$10, $39, $100	Monthly GitHub AI Credits
Codex	$8, $20, from $100	Included with ChatGPT plans

For the full breakdowns, see Claude Code vs Cursor, Claude Code vs Codex and Claude vs Copilot.

Which option to choose

• Trying it, or one person building client tools: start with Pro. Move to Max if you hit limits often.
• A small agency team: Team seats keep billing and admin in one place. Give heavy users Premium seats.
• Automation, CI and agents you build for clients: use an API key. Anthropic doesn't allow claude.ai logins in products built on its Agent SDK, so client-facing agents use API billing (Agent SDK tutorial).

For client work, put model usage in your pricing. It's a real cost of delivery, like SMS or hosting (how to price it).

Is it worth it?

Measure it against the work. If it saves you a few hours a month on client builds, a $20 plan usually pays for itself, but that depends on your work and rates. Start with Pro, check /usage for a month, and move up only if limits slow you down.

Learn to build with Claude Code

The Stride AI Coding course teaches Claude Code and Codex for building tools and apps for clients, written for non-programmers, with weekly live calls. Many of the free guides cover Claude Code setups.

FAQ

How much is Claude Code a month?

It costs $20 a month on Pro, or $17 a month billed yearly. Max is $100, Ultra is $200, and Team seats are $25 or $125 billed monthly. There's no separate Claude Code price: it comes with those plans.

Is Claude Code free?

No. Claude's Free plan doesn't include it. It comes with the paid Pro, Max, Team and Enterprise plans, or you can use it with an API key and pay per token through the Claude Console.

Does Claude Code cost per token?

Only with an API key. On a subscription you pay a flat monthly price and get usage limits. If you turn on usage credits, anything past your limit bills at standard API rates.

Is the $20 Claude plan enough for Claude Code?

For many individuals, yes, especially if you clear sessions between tasks and use Sonnet for most work. If you hit usage limits regularly, Max is $100 a month, or you can turn on usage credits to keep working past the limit.

How much does Claude Code cost per month on the API?

It depends on how you use it. Anthropic's docs report an average of $150 to $250 per developer per month across enterprise deployments, and under $30 per active day for 90% of users. Run a pilot to measure your own.

How do I see how much I've used?

Type /usage in a session. Subscribers see plan usage bars and a breakdown of what used them. API users see the session's tokens and an estimated cost. The Claude Console's Usage page is the authoritative bill.