Two tools that now meet in the middle
Anthropic's overview describes Claude Code as "an agentic coding tool that reads your codebase, edits files, runs commands, and integrates with your development tools." The full-featured version is the command-line tool, and the same engine runs in VS Code, JetBrains IDEs, a desktop app and on the web, including from the Claude apps for iOS and Android. Anthropic's help center lists Cursor and other VS Code forks among the supported editors, with usage drawn from the same Claude plan.
Cursor is an editor first. Its Agent sits in a side panel and, in Cursor's words, "can complete complex coding tasks independently, run terminal commands, and edit code." Cursor also has a terminal CLI with Agent, Plan and Ask modes, and cloud agents that run in isolated virtual machines in the cloud and can be started from the desktop app, the web, the iOS app, Slack, GitHub, Linear or an API.
What each subscription costs
Anthropic doesn't include Claude Code on its Free plan. Cursor's lineup is sold through The X Subscription, one subscription for Grok, Cursor and X.
| Claude plan | Monthly price | Claude Code |
|---|---|---|
| Free | $0 | Not included |
| Pro | $20, or $17 a month billed annually at $200 | Included |
| Max 5x | $100, monthly only | 5x Pro's usage per five-hour session |
| Max 20x | $200, monthly only | 20x Pro's usage per five-hour session |
| Team Standard seat | $25, or $20 billed annually | "More usage than Pro" |
| Team Premium seat | $125, or $100 billed annually | 5x a Standard seat |
| Cursor plan | Monthly price | Included usage |
|---|---|---|
| Free | $0 | Limited agent and chat requests |
| Plus | $8, or $80 a year | Base pool, Grok models only |
| Premium | $30, or $300 a year | 4x Plus |
| Super | $100, or $1,000 a year | 15x Plus |
| Ultra | $200, or $2,000 a year | 50x Plus |
| Teams | $30, $100 or $200 per seat | Matches the seat's tier |
The Max 20x price comes from Anthropic's help center; the pricing page itself says Max starts "From $100." Anthropic's pricing page also now lists $100 or $200 a month in API credits for building on the Claude Platform with Max 5x and Max 20x. Enterprise is $20 per seat per month, billed annually with a 20-seat minimum, plus usage at API rates. Cursor's Enterprise pricing is custom.
Anyone still on an older Cursor plan, such as Pro at $20 or Pro+ at $60, can keep it. Cursor says each new paid plan includes "at least twice as much coding usage with Cursor models" as the legacy plan it replaces, but drops the old plans' included allowance for third-party models.
Which models your money buys
This is where the two diverge most. On Pro, Max, Team and Enterprise, Claude Code's default model is Claude Opus 5.5, at medium effort, according to Anthropic's model configuration docs. You can switch to Sonnet 5.5 or Haiku 5.5, and Fable 5.1 is available too. Anthropic's pricing page says Fable runs only on usage credits on Pro, while Max can spend up to half of its weekly limits on it. Anthropic's pricing page calls Opus 5.5 a "Daily driver for agentic coding and enterprise work."
Cursor offers a wider catalog. Its documentation lists Claude Fable 5.1, Opus 5.5, Sonnet 5.5 and Haiku 5.5, Gemini 4 Argon and Gemini 3.1 Pro, GPT-5.6 Sol, Terra and Luna, Muse Wonder, and Cursor's own Grok 4.7, Grok 4.6, Grok 4.5 and Composer 2.5. Only the last four draw from the weekly pool your plan includes. Third-party models, Claude among them, are available from Premium up and billed at API list price from prepaid credit starting with the first request. Plus doesn't include them at all.
Cursor publishes those list prices per million tokens. Claude Opus 5.5 costs $4 for input and $20 for output in Cursor, and Claude Sonnet 5.5 costs $2 and $10. On Teams and Enterprise, eligible third-party requests also carry a Cursor Token Rate of $0.25 per million tokens, which Cursor says doesn't apply to individual plans.
How usage is metered
Anthropic meters Claude Code with two limits at once, according to its help center: a session limit that resets every five hours and a weekly limit. Both are shared with Claude on the web, desktop and mobile, and Anthropic publishes multiples rather than amounts. Our guide to how Claude Code usage limits work covers the details. When the five-hour limit hits mid-response, Anthropic is rolling out a small wrap-up allowance that lets Claude Code finish the step in progress, up to once per weekly period on Pro and each time on Max. Beyond that, paid plans can turn on usage credits.
Cursor uses a single weekly pool instead, also described only in multiples of Plus. Cursor's pricing page says the pool is "shared across Cursor and Grok," and its help center adds Grok Bot and the Grok apps to the list. When the pool runs out, Cursor says you can wait for the weekly reset, keep going on prepaid credit at the same per-model rates, or upgrade, and that "requests are never downgraded in quality or speed." Prepaid credit doesn't reset weekly and currently expires 365 days after purchase.
Both companies share your allowance with a chat product, so a lot of Claude chat or Grok use leaves less for coding.
Agent features side by side
The feature lists have converged. Claude Code reads project instructions from a CLAUDE.md file, or from an existing AGENTS.md, and supports skills, hooks and MCP servers. Anthropic says a lead agent can coordinate several Claude Code agents on parts of one task. Cloud scheduled tasks keep running when your computer is off and can be triggered by GitHub events, and Claude Code also works with GitHub Actions, GitLab CI/CD, automatic GitHub code review and Slack.
Cursor's Agent can search the codebase and the web, edit files, run shell commands, drive a browser and take checkpoints you can roll back to. Projects let a coordinator agent hand work to other agents. Cloud agents, Automations, Bugbot code review and Origin code hosting start at Premium. Recent changelog entries added remote control of local agents from the iOS app on October 6 and an Origin app marketplace on October 13.
On independent measurement, the comparison is lopsided. Artificial Analysis's Coding Agent Index includes several Claude Code setups, with Claude Sonnet 5.5 (max) tied at the top at 68.4, as our ranking of the best coding agents shows. Cursor's own agent doesn't appear on the leaderboard.
Which one to pick
If you mainly want Claude, Claude Code is the more direct route. A $20 Pro plan covers Opus 5.5 in the terminal and in your editor, Cursor included, and the same plan covers Claude chat. In Cursor, every Claude request is paid by the token on top of a Premium plan or higher.
Cursor suits anyone who wants an editor built around the agent and a choice of models from several companies, and is happy to rely on Cursor's own Grok and Composer models for the included usage. Its $8 Plus plan is the cheapest paid way into either tool, though it is limited to those models. Cursor also bundles Grok and X benefits, which add value only if you use Grok or X.
Heavy users hit similar price ceilings, with both top individual tiers at $200 a month, but the money buys different things: more Claude usage per session on Max, or a larger weekly pool of Cursor models on Ultra. For how Anthropic's plans stack up against OpenAI's, see our Claude Code vs Codex comparison. Both companies change these plans often, and we'll update this page as the plans change.