Skip to content
Neural Minute

Review · Build and data

Claude Code Review 2026: What $20 a Month Unlocks

By Maxime Aulne, Founder and Editor

Claude Code comes with any paid Claude plan from $20 a month. What the terminal agent does, where the five-hour limit bites, and who should pay.

Minute Verdict

Claude Code is Anthropic’s coding agent for the terminal: it reads a repository, edits files across several of them, runs the tests and opens the pull request. It has no price of its own, because it ships inside every paid Claude plan. The cheapest seat is Pro at $20 a month, or $17 a month with $200 paid up front for the year.

Buy Pro if you already pay for Claude and want the coding agent on the same bill, since the terminal tool is included rather than sold separately. Stay away if you need a number before you commit: the agent draws on the same five-hour and weekly budget as the chat window, and no message count is published. The Claude review covers what the rest of that subscription opens.

Best for: a solo developer or a small team that already works in Claude and wants a coding agent without a second invoice.

Not for: a buyer who needs a predictable monthly volume, or anyone who wants to try before paying, since the Free plan carries no Claude Code.

From: $20/mo billed monthly, checked November 16, 2026

What Claude Code does

Claude Code is an agentic coding tool rather than an autocomplete layer. You describe the job in plain language, it plans the approach, then it reads the relevant files, edits them, runs commands and checks the result. Anthropic ships the same engine on six surfaces: a terminal CLI, a VS Code extension that also installs in Cursor and Devin Desktop, a JetBrains plugin, a desktop app, a browser session at claude.ai/code, and Slack. GitHub Actions and GitLab CI/CD put it in a pipeline as well.

The surfaces are not separate products. A CLAUDE.md file at the project root, which carries coding standards and review checklists, is read at the start of every session on every surface, and MCP servers and settings follow the same rule. The desktop app bundles the CLI, so nothing has to be installed beside it, while the JetBrains plugin requires the CLI installed separately. The browser version needs no local setup, which suits a repo you do not have on the machine.

What it does well is the long job. Anthropic describes it following imports across a repository, running the test suite and continuing when something breaks, and coming back to you when it needs a decision. The composable side is the part that changes a working day: you can pipe a log into it, run it non-interactively in CI with claude -p, or chain it with git, which is how a team turns a repeated review into one command. It is not an editor, so the visual side of a change still happens in your IDE, and anyone shopping on that ground should read our Cursor CLI review as the other terminal option.

On this page

What you get for the money

Plan Price What is included The limit that matters
Free $0 Chat, files, code execution, memory, up to 5 projects No Claude Code, no Opus
Pro $20/mo, or $17/mo with $200 a year up front Claude Code, Projects, Research, Skills, Claude in Chrome Shared 5-hour and weekly pool, Fable only through usage credits
Max $100/mo 5x Pro usage, Fable up to half the weekly limit, $100 a month in API credits Billed monthly, one seat
Ultra $200/mo 10x Pro weekly usage, no 5-hour limit on verified accounts, $200 a month in API credits Billed monthly, one seat
Team Standard $25/seat/mo, or $20 billed annually Every Claude feature, more usage than Pro 2 to 150 seats, no invoicing below Enterprise
Team Premium $125/seat/mo, or $100 billed annually 5x the usage of a Standard seat The expensive seat is the only one with real volume
Enterprise $20/seat/mo billed annually SCIM, audit logs, spend limits, role-based access Usage billed at API rates on top of the seat

The price buys one pool, not one product. Pro at $20 covers the chat window, long documents, Research, Skills and the coding agent, and the annual view drops the same seat to $17 a month. A year of Pro costs $240 at the monthly rate, or $200 paid up front.

The second route is a Console account, billed per token on prepaid credits, which is what the CLI uses when you log in with an API key instead of a subscription. Anthropic’s cost guide puts enterprise deployments at about $13 per developer per active day and $150 to $250 per developer per month, with 90 percent of users staying below $30 on an active day. Those figures are Anthropic’s own and shift with the repository and the models chosen.

Where it falls short

  • The Free plan does not include Claude Code. Trying the agent therefore costs $20, and there is no standalone coding plan for a developer who wants the terminal tool without the chat subscription.
  • Usage is sold as a five-hour session window plus a weekly cap, with no message count. The five hours start at your first message, so a heavy afternoon can end the session before the work is done.
  • The pool is shared. Chat on the web, the desktop app, mobile and Claude Code all draw from the same budget, which means a coding session spends what you would otherwise use for the rest of the day.
  • When the budget runs out, the choices are to wait for the reset, move to Max or Ultra, or turn on usage credits billed at standard API rates. The last option is honest and useful, and it also turns a flat fee into a bill you cannot predict.
  • It is not open source. The GitHub repository collects issues, plugins and the changelog, the licence reads copyright Anthropic, all rights reserved, and the npm package carries the same restriction.
  • The terminal is the first surface and it shows. Installation is a shell script, sign-in happens in a browser, and the Windows path mixes PowerShell and CMD commands, which is where the install errors come from.

Claude Code alternatives

  • Cursor, from $8 a month on the Plus seat of the X Subscription, is an editor first, with its own agent, a model picker and a free tier, and the $30 Premium seat is where cloud agents and pull request review start. Cursor review
  • Grok Build, from $8 a month, is xAI’s terminal agent, with plan mode, subagents and a small free allowance of agent requests before you pay. Grok Build review
  • Kimi Code, from $19 a month, is a terminal agent sold inside the Kimi membership, and it imports Claude Code sessions from the last 15 or 30 days, so a repository does not start from zero. Kimi Code review

Who it is for: Claude Code

It fits a developer who already pays for Claude and codes in a terminal. The agent costs nothing extra on Pro, it reads the projects and instructions you have already set up, and it works on the same repository from the CLI, VS Code or the browser. A small team gets the same deal on Team seats, with central billing and a pooled API credit of up to $500 a month.

It fits less well when the buyer wants a coding tool with its own price. A team that needs a per-seat spend cap it can forecast, or a developer who only wants the terminal agent, is looking at Cursor, Grok Build or Kimi Code instead. The rest of that field sits in the build tools hub. The exception is a company that already runs Claude across other teams: Enterprise adds SCIM, audit logs and spend limits, and the seat rate is $20 a month before usage.

How we checked

We read the pricing page in its monthly and annual views, the Claude Code product page, the official documentation for the overview, the quickstart and the cost model, the help center article on usage limits, and the GitHub repository with its licence, all on November 16, 2026. We have not used Claude Code yet, so this is a research review, and no capture or result on this page comes from a session of ours. A hands-on test would answer the one question the documents leave open: how many real tasks one $20 session budget buys on a repository of a given size. The label turns to Hands-on the day that test happens.

What to do next

If you already pay for Claude, install the CLI with the native script, run it once in a small repository and write a short CLAUDE.md before the first real task, since that file is what keeps the agent inside your conventions. If you do not pay for Claude yet, take Pro for one month at $20 and watch Settings then Usage for the first week: the reset pattern tells you whether Pro is enough. A developer who codes in it all day ends up on Max at $100, and anyone who wants the terminal agent alone should compare the alternatives above first.

FAQ

Is Claude Code free? No. The Free plan does not include Claude Code, and no free trial is published. It is included from Pro up, at $20 a month billed monthly.

How much does Claude Code cost? Nothing extra on top of a Claude plan. Pro is $20 a month, or $17 with $200 paid up front, Max is $100, Ultra is $200, Team seats are $25 and $125 billed monthly, and Enterprise is $20 a seat a month billed annually plus usage at API rates. A Console account is the other route, billed per token.

Is Claude Code open source? No. The repository at github.com/anthropics/claude-code collects issues, plugins and the changelog, and its licence reads copyright Anthropic, all rights reserved, use subject to the commercial terms. The npm package carries the same restriction and asks for Node 22 or newer.

Does Claude Code need a Claude subscription? Not always. The terminal, the VS Code extension and the JetBrains plugin also accept a Console API key and third-party providers. The desktop app, the browser version and the mobile apps need a paid Claude plan.

What happens when I hit the usage limit? The session resets five hours after your first message, and the weekly allowance resets on a fixed day. You can also upgrade to Max or Ultra, or turn on usage credits on a paid plan, which bills extra work at standard API rates.

Sources

Sources

Update log

Published November 16, 2026, prices checked the same day on claude.com/pricing in both billing views.

Newsletter

One email when a review lands.

That is when the promo codes are still fresh, before the prices move.

Nothing else. Unsubscribe in one click.