Skip to content
Mia Automation
NLEN Sign in Put Mia to workGet started

Is Claude Code free? What it costs and the cheapest ways to use it

No. The Claude Code program is free to install, but using it needs a paid Claude plan or API credit, and the free Claude plan does not include it. The cheapest plan is Pro at $20 a month, or about $17 a month billed yearly. On the API a small coding task cost us about $0.09 with Sonnet 5.5, and a spend cap stops a run at the amount you choose.

Pricing and comparisonsUpdated 6 min read

Tested by us on 11 November 2026.

The short answer: free to install, not free to use

Anthropic does not charge for the Claude Code program. The installer, updates and the VS Code extension cost nothing. The model behind it does cost money, and you pay for it in one of two ways: a paid Claude plan or API credit on a Console account. The free Claude plan does not include Claude Code; the comparison table on claude.com/pricing marks it “No” for Free (checked November 11, 2026).

Feature comparison table on the Claude pricing page: the Claude Code row shows a cross under Free and a check mark under Pro, Max and Ultra (opens at full size)
The plan comparison on claude.com/pricing, November 11, 2026: Claude Code is not part of Free.
Prices from claude.com/pricing, checked November 11, 2026, excluding tax. Cost per task from our own test below.
RouteWhat you payWhat you give up
Free Claude plan$0Claude Code is not included
Pro, billed yearly$200 a year, about $17 a monthShared usage pool with chat, five-hour and weekly limits
Pro, billed monthly$20 a monthSame limits; cancel any month
API credit with a spend capPer token; about $0.09 per small coding task with Sonnet 5.5 in our testNo fixed price, so set a cap yourself
Another agent with a local modelYour own hardware and electricityNot Claude Code, and Anthropic does not support it

Route 1: Pro is the cheapest plan that includes Claude Code

Pro costs $20 a month, or $200 up front for a year, which works out to about $17 a month. You get Claude Code in the terminal and in your editor, plus the chat app, from one usage pool. That pool resets every five hours and has a weekly cap on top. When you reach it, you wait, upgrade or switch on usage credits, which bill at API rates.

For most people who try Claude Code a few times a week, Pro is the simplest choice. The price is fixed, and the limit protects you from a surprise bill. Our Claude Code pricing page compares Pro with Max and Ultra.

Route 2: pay per task with a daily budget

Without a subscription, you sign in with a Claude Console account and pay for the tokens you use. To see what that costs, we ran five ordinary coding tasks on a small open-source Python library. We ran the set three times per model. Sonnet 5.5 needed $0.41 to $0.47 for all five at API list price, about $0.09 per task. Opus 5.5 needed $1.37 to $1.46, about $0.28 per task. These were short tasks in fresh sessions; a long session with many follow-ups costs two to three times as much per task.

Our test, November 10, 2026: Claude Code 2.1.321, five tasks on python-slugify. Daily budgets are our own calculation from the measured averages.
Daily budgetSonnet 5.5 tasks per dayOpus 5.5 tasks per dayPer month (21 working days)
$0.50about 5about 2$10.50
$1about 11about 3$21
$3about 35about 10$63

Two caps keep this under control. For a single unattended run there is --max-budget-usd. We set it to $0.10 on a feature task with Opus 5.5. Claude Code stopped after three turns at $0.105 and reported Reached maximum budget ($0.1), without having changed a file. For your account as a whole, set a spend limit on the workspace in the Claude Console.

Animated diagram: three turns of a Claude Code run, a running total that passes the $0.10 cap, and the run stopping at $0.105 without changing files
How the spend cap behaved in our test: it checks between turns, so the run ended slightly above the cap.
bash
# one run, stopped at ten cents
claude -p "Fix the failing test in tests/test_slug.py" --model sonnet --max-budget-usd 0.10

Print mode (-p) only. The cap is checked between turns, so a run can end slightly above it.

At around $1 a day the API is about as expensive as Pro, without the five-hour window. Below that, it is cheaper. Above that, Pro usually wins.

Route 3: a free or local model, and what you lose

You can point Claude Code at another address with ANTHROPIC_BASE_URL, and some gateways translate requests to other models. Anthropic's gateway documentation says plainly that it does not support routing Claude Code to non-Claude models through any gateway (checked November 11, 2026). Features can break with each update, and you are on your own when they do.

If you want an agent that runs on a free or local model, pick one that is built for it. OpenCode, for example, finds models served by Ollama or LM Studio on your own machine, according to its provider documentation. You give up model quality on hard tasks and you need a computer that can run the model. In return no code leaves your machine.

Claude Code or the Claude app: what is the difference?

The Claude app is a chat on the web, desktop and phone. You type, Claude answers, and you copy the result. Claude Code is an agent that works inside a project folder: it reads your files, edits them and runs commands such as your test suite, asking before risky steps. Both draw from the same usage pool on a paid plan. Only the app is available on Free.

Questions

Is Claude Code free?

The program is free to install, but using it is not. You need a paid Claude plan (Pro, Max, Ultra, Team or Enterprise) or API credit on a Console account. The free Claude plan does not include Claude Code.

What is the cheapest plan that works with Claude Code?

Pro: $20 a month, or $200 for a year (about $17 a month). On the API you can spend less, about $0.09 per small task with Sonnet 5.5 in our test.

What does Claude Code cost per day on the API, and how do I set a budget?

In our test a small coding task cost $0.09 with Sonnet 5.5 and $0.28 with Opus 5.5. Use --max-budget-usd to stop a single run and a workspace spend limit in the Claude Console to cap the account.

Can I use Claude Code with a free or local model?

Technically some gateways allow it, but Anthropic does not support non-Claude models in Claude Code. For a local model, use an agent built for it, such as OpenCode with Ollama.

Is there a free trial of Claude Code?

We found no free trial for individuals on the pricing page on November 11, 2026. Trials are listed only for sales-assisted Enterprise plans.

What is the difference between Claude Code and the Claude app?

The app is a chat. Claude Code works inside your project: it reads and edits files and runs commands. Both share one usage pool on paid plans.

Sources