Last updated: 2026-11-02
What changed
Claude Code is Anthropic's agentic coding tool, first released in February 2025, that runs in the terminal, IDEs, a desktop app, the web and mobile. It reads a codebase, edits files across a project, runs tests and opens pull requests. It comes with the Pro, Max, Ultra, Team and Enterprise plans, and has no free plan.
About Claude Code
Built by Anthropic, this coding agent reads a codebase, plans a change across files, edits them, runs the tests, iterates on failures, then commits and opens the pull request. It began in February 2025 as a terminal tool and now also runs in VS Code (and Cursor), JetBrains IDEs, a desktop app for macOS and Windows with a Linux beta, the web, the Claude iOS and Android apps, Slack and GitHub Actions, according to Anthropic's documentation. Anthropic raised a $65 billion Series H at a $965 billion post-money valuation in 2026, reported by law firm Orrick, and filed an S-1 for a proposed initial public offering on 26 October 2026.
Recent changes are mostly about running longer with less supervision. Auto mode is the default on Pro, Max and Team plans, self-hosted environments are in public beta so sessions can run inside your own network, and Projects groups related sessions so several agents can be supervised from one screen. Recurring jobs can also be put on a schedule. Anthropic's pricing page lists Opus, Sonnet and Haiku on the paid plans, with Fable available through usage credits on Pro or a share of weekly limits on Max and Ultra, and context windows of up to 1 million tokens depending on the model.
The individual ladder changed shape this autumn: the two old Max tiers became Max and Ultra, and both now include a monthly allowance of API credits for building on the Claude Platform. Ultra is the plan for long unattended runs, since it drops the 5-hour usage window for verified accounts.
Claude Code competes in the AI coding assistants category with Cursor, Windsurf, GitHub Copilot, Aider, plus the hosted engineer Devin. For a head-to-head read, see Cursor Composer 2 vs Claude Code and the best AI coding assistants of 2026. The general Claude assistant page covers the chat product the same plans unlock.
Where the alternatives win: Claude Code only runs Anthropic's models, so a team with a multi-vendor procurement policy will find Cursor easier to approve, and a developer who wants suggestions and visual diffs while typing will do less work in an editor-first tool than in a terminal agent. It also has no free plan, so someone who needs occasional help is better served by a free-tier assistant. Teams that run it at scale should plan around usage limits, which Anthropic describes as multiples of the Pro allowance rather than fixed token counts.
Pricing
No free tier: the Free plan does not include Claude Code. Pro is $20/month ($17/month billed annually, $200 up front). Max is $100/month with 5x Pro usage and $100 a month in API credits; Ultra is $200/month with no 5-hour limit for verified accounts, 10x Pro usage a week and $200 a month in API credits.
Max and Ultra are billed monthly only. Team Standard is $25/seat/month ($20 annual) and Team Premium is $125/seat/month ($100 annual), for teams of 2 to 150, with up to $500 a month in pooled API credits. Enterprise is $20/seat/month billed annually with a 20-seat minimum, plus usage at API rates.
Prices exclude tax and Anthropic can change them. Checked against Anthropic's pricing page on 2026-11-02. Before choosing, compare the per-user plans of Cursor.
| Tier | Monthly price | What it includes |
|---|---|---|
| Pro | $20/mo | Claude Code included, More usage than the Free plan, Opus, Sonnet and Haiku, Fable through usage credits |
| Enterprise | $20/mo | Claude Code included, Billed annually per seat, 20-seat minimum, usage at API rates, SCIM, audit logs and a Compliance API, HIPAA-ready offering available |
| Team Standard | $25/mo | Claude Code and Claude Cowork included, More usage than Pro, Central billing, SSO and admin controls, No model training on your content by default |
| Max | $100/mo | Claude Code included, 5x Pro usage every 5 hours and every week, Fable up to half the weekly limit, Monthly API credits for the Claude Platform, Priority access at high traffic times |
| Team Premium | $125/mo | Claude Code included, Five times the usage of a Team Standard seat, Seat types can be mixed in one team, Pooled API credits shared across the team |
| Ultra | $200/mo | Claude Code included, Verified accounts skip the 5-hour window, A weekly allowance ten times the size of Pro, Double the API credits of Max |
Feature Comparison by Tier
| Feature | Pro | Max | Ultra | Enterprise | Team Standard | Team Premium |
|---|---|---|---|---|---|---|
| Monthly price | $20 | $100 | $200 | $20/seat + usage | $25/seat | $125/seat |
| Usage allowance | Baseline | 5x Pro | 10x Pro weekly, no 5-hour cap | API rates | More than Pro | 5x Team Standard |
| Claude Code included | Yes | Yes | Yes | Yes | Yes | Yes |
| Monthly API credits | - | $100 | $200 | - | Pooled, up to $500 per team | Pooled, up to $500 per team |
| Central billing and SSO | - | - | - | Yes | Yes | Yes |
| Audit logs and SCIM | - | - | - | Yes | - | - |
| Priority access at high traffic | - | Yes | Yes | - | - | - |
Key Features
- Agentic coding loop: Claude Code plans a task, reads the files it needs, edits across the project, runs the tests and opens a pull request through GitHub, GitLab or your own command line tools.
- Nine surfaces, one engine: The terminal, VS Code, JetBrains, the desktop app, the web, iOS, Android, Slack and GitHub Actions all run Claude Code, so a team can mix a CLI habit with a visual diff review.
- Auto mode by default: Since 7 August 2026 Pro, Max and Team sessions start in auto mode, which lets Claude work longer between prompts while still catching risky commands.
- Self-hosted environments: In public beta since 6 August 2026, sessions can run on your own infrastructure inside your network, next to internal services that a cloud session cannot reach.
- Projects for parallel agents: Projects, released in September 2026 on the desktop app, group related sessions so you can supervise several Claude agents at once from a single view.
- MCP, skills and hooks: The open Model Context Protocol connects Claude Code to Jira, Slack and Google Drive, while CLAUDE.md, skills and hooks encode a team's own standards and checks.
- Scheduled tasks: Recurring jobs such as a morning PR review run in the cloud, where they survive a closed laptop and can trigger on GitHub events, or on your own machine with local file access.
Pros
- One engine across the terminal, VS Code, JetBrains, desktop, web, mobile, Slack and GitHub Actions, so you pick the surface that suits the task.
- Sessions start in auto mode on Pro, Max and Team, so long tasks stop less often to ask for permission.
- Self-hosted environments (public beta) let sessions run inside your own network, which matters when the code cannot leave it.
- Sub-agents and background agents run several sessions side by side, and Projects groups them for supervision.
- CLAUDE.md, skills, hooks and MCP servers let a team write its standards into the tool instead of repeating them in prompts.
Cons
- No free plan: the Free tier excludes Claude Code, so the first step is a paid Pro seat or a metered Console account.
- Anthropic models only: there is no way to route a task through a non-Claude model, which rules it out for strict multi-vendor policies.
- Usage limits are described as multiples of the Pro allowance, not fixed token counts, so heavy users cannot predict when a window will run out.
- Anthropic documents extensions only for VS Code and JetBrains; developers on other editors use the terminal CLI.
Product Information
- Cloud
- Yes
- Self-Hosted
- No
- On-Premise
- No
- Languages
- English
- Training
- Documentation, Community Discord, Video Tutorials
Data Handling
- Training-data policy
- Free, Pro, Max and Ultra users can opt out of model training; Team and Enterprise plans do not train on your content by default (Anthropic pricing page, 2026-11-02).
- Data retention
- 30 days
- Compliance
- SOC 2 Type II · ISO 27001:2022 · ISO/IEC 42001:2023 · GDPR DPA
Frequently Asked Questions
How much is Claude Code a month on each plan?
Claude Code comes inside paid Claude plans rather than as a separate product. Pro costs $20 a month or $17 on annual billing, Max is $100 and Ultra is $200, and both of those add monthly API credits. Team seats are $25 Standard or $125 Premium on monthly billing, and Enterprise is $20 a seat billed annually plus usage at API rates.
Can I use Claude Code without paying?
Not on a subscription. Anthropic's plan table marks Claude Code as unavailable on Free, so the cheapest route is Pro, or a Claude Console account billed per token. The free chat plan still lets you try the underlying models on code questions, without the agent that edits files and runs commands.
What are the main Claude Code alternatives?
Cursor and Windsurf are editor-first agents, GitHub Copilot is built around GitHub, Aider is an open-source terminal tool and Devin is a hosted autonomous engineer. They cover the same job from different angles, and the AI coding assistants directory lines them up on price and platform.
Is Cursor a better choice than Claude Code?
It is when you want to stay in an editor and accept changes line by line as you type, or when policy requires more than one model provider. Claude Code is the stronger fit for handing over a whole job, such as a dependency upgrade, and reviewing the finished pull request afterwards.
What is the difference between Claude Max and Ultra for coding?
Max gives five times Pro usage in each 5-hour window and each week. Ultra removes the 5-hour window for verified accounts, raises the weekly allowance to ten times Pro and doubles the monthly API credits. Both cap Fable at half the weekly limit, so the choice comes down to how long your sessions run.
Top Alternatives
- Cursor: Choose Claude Code to hand off whole tasks from a terminal agent; pick Cursor for an editor-first workflow with a choice of model providers.
- Windsurf: Choose Claude Code for terminal-driven agent work across many surfaces; pick Windsurf for an editor-first surface.
- GitHub Copilot: Choose Claude Code to delegate whole tasks from a terminal agent; pick GitHub Copilot for a free plan inside the IDEs and GitHub workflow you already use.
Hokai guides covering Claude Code
- Cursor Composer 2.5 vs. Claude Code: Which AI Coding Agent Should You Use in 2026?: Cursor Composer 2.5 vs Claude Code on October 2026 pricing: Cursor's new Grok-bundled plans, Claude's 5.5 models, token costs and who should pick which.
- Claude Max Limits After the 50% Boost: What Changed in 2026: Anthropic ended the 50% Claude Code weekly boost on 13 Sept and kept a 25% rise. Claude Max costs $100 or $200 a month. Current limits and what overflow costs.
- Oracle Added Claude Code to Fusion. Its Analyst Ranking Didn't Move.: Oracle added Fusion Claw and 25 agent apps on 29 September, yet Futurum's Signal report still rates Oracle in its lowest platform zone. What it means.
- Codex vs Claude Code in 2026: Which Coding Agent Should You Run?: Codex vs Claude Code in 2026: independent scores, cost per task, plans, usage limits, cloud agents and security scanners compared, with a pick per team type.
Also mentioned in
- ChatGPT vs Claude in 2026: Which One Should You Pay For?: ChatGPT vs Claude on 5 Nov 2026: every plan from $8 to $500, which model each one gives you, and a winner for each job, read from the vendors' own pages.
- Gemini 4 Argon vs GPT-6 Astra vs Claude Opus 5.5: Cost Per Real Task: Argon, GPT-6 Astra and Claude Opus 5.5 compared on what a finished task costs, not the token price, plus the 1 December price rise on Argon. Updated Oct 2026.