The datasheet for every AI agent

Coding agents

Visit Cursor Compare

The AI-first code editor (a VS Code fork) with a coding agent that reads, plans and edits across your whole codebase. Free Hobby tier, Pro from $20/month.

From
Free
Pricing model
freemium
Free tier
Yes
Deployment
Cloud
Interface
IDE
Open source
No
Model / LLM
Anthropic Claude, OpenAI GPT, Go…

Our verdict

Cursor remains the best in-editor agent for developers who want AI woven into a familiar editor rather than bolted onto a separate tool.

Its multi-file agent mode and tab-completion are the most polished in the category, and the model choice is wider than competitors: Anthropic, OpenAI, Google and, since the SpaceX close, Cursor's own Grok 4.6 (released 2026-08-12).

A new Cursor Router, generally available since 2026-07-22, automatically classifies each request and dispatches it to the right model, with Cursor's own online A/B tests showing frontier-quality output at roughly 60% lower cost than sending every request to the strongest model. The tradeoffs are unchanged from the last refresh.

Cursor is closed-source and subscription-priced, heavy daily use can push buyers from Pro at $20/month up to Ultra at $200/month, and post-acquisition ownership now sits with SpaceX, which also owns the rival xAI brand. For buyers who value per-request choice, none of this changes the day-to-day editor. For buyers whose procurement requires multi-vendor neutrality on a 12-18 month horizon, watch the consolidation.

What is Cursor?

Cursor is an AI-first code editor, a fork of VS Code, built by Anysphere around an in-editor coding agent. It pairs predictive tab-completion with an agent that reads the whole repository, plans a change and applies edits across many files. Cursor now ships a CLI, a web management surface, an iOS app, cloud agents that keep working after the laptop closes, and a Google Workspace plugin.

As of 2026-08-14, Cursor's parent company Anysphere is a part of SpaceX, with the Cursor team joining SpaceXAI to train and ship its own models alongside Claude, GPT and Gemini. Cursor's published ARR doubled to roughly $2 billion between December 2025 and March 2026 (Bloomberg, 2026-03-02), and on 2026-05-22 Gartner named it a Leader in the 2026 Magic Quadrant for Enterprise AI Coding Agents.

What does Cursor do?

You describe a change in natural language and Cursor's agent gathers the relevant files across the repo, proposes a multi-file edit and applies it as a reviewable diff. Beyond the desktop editor, Cursor ships a CLI, a web app for managing agents, an iOS app, a Google Workspace plugin, and cloud agents that run long-running tasks asynchronously after you close the laptop.

Cursor Router, generally available for Teams and Enterprise since 2026-07-22, classifies each request and routes it to the most capable model for the task. Bugbot now bills on usage-based metering rather than a flat seat. The platform runs on Anthropic Claude, OpenAI GPT, Google Gemini and Cursor's own Grok 4.6, all selectable per request, plus predictive tab autocomplete and codebase-wide chat with retrieval.

How Cursor works

  1. Describe a change in natural language; the agent gathers the relevant files across the repo.
  2. It proposes a multi-file edit and applies it as a diff you review before accepting.
  3. Predictive tab autocomplete suggests the next edit as you type, not just the next token.
  4. Cursor Router classifies each request and routes it to the most capable model for that task, with separate Intelligence, Balance and Cost modes for teams and admins.
  5. Hand a longer task to a cloud agent that runs asynchronously across desktop, web, CLI or the iOS app.

Key features

Agentic multi-file edits
Plans and applies changes across many files from one instruction, with a reviewable diff before accept.
Predictive autocomplete
Tab-completion that suggests the next edit, not just the next token.
Codebase chat with retrieval
Ask questions across the whole repo and the editor retrieves the relevant context to answer.
Cursor Router
Generally available for Teams and Enterprise since 2026-07-22. Classifies each request and routes it to the best model, with Cursor reporting roughly 60% cost savings versus sending everything to a frontier model.
Grok 4.6 in-app
Cursor's own Grok 4.6 model, released 2026-08-12 jointly with SpaceXAI, available inside Cursor for long-running agent work and interactive builds.
Cloud agents with background runs
Hand a task to a cloud agent that keeps working after the editor is closed, with team-shared context for collaborative runs.

What are Cursor's use cases?

Ship features faster
A developer describes a feature end-to-end and the agent handles the cross-file plumbing, with the multi-file diff as the review surface.
Cross-codebase refactors
Rename and restructure across a large codebase with the agent tracking every call site.
Hand off long-running work
Kick off a cloud agent that keeps running after the laptop closes and surfaces results back into the editor for review.
Spend less on frontier models
Teams turn on Cursor Router and let it send routine edits to cheaper models while keeping hard reasoning on the frontier, with Cursor reporting 60% cost savings against all-frontier routing.

Who is Cursor for?

  • Professional developers who want a powerful AI agent woven into a familiar VS Code-based editor
  • Teams doing large refactors who want the agent to track every call site across the codebase
  • Engineering organisations that want Claude, GPT, Gemini and Cursor's own Grok on one subscription without separate accounts
  • Teams that want a router to send routine work to cheaper models and reserve frontier pricing for the hard tasks

Not forTeams that want a terminal-only, fully scriptable agent to drive from CI, or buyers whose procurement requires long-term multi-vendor neutrality on a 12-18 month horizon now that Cursor's parent Anysphere is part of SpaceX (owner of the rival xAI/Grok models).

What does Cursor integrate with?

  • GitHub & Git
  • Model providers: Anthropic, OpenAI, Google, SpaceXAI Grok 4.6
  • MCP servers and tools
  • Google Workspace plugin (added August 2026)
  • VS Code extension ecosystem (it is a fork)
  • iOS app for kicking off and reviewing agent runs
  • Web app for managing agents across desktop, web and mobile

Why use Cursor?

  1. The most polished in-editor agent experience for day-to-day coding.
  2. Wider model roster than competitors, with Anthropic, OpenAI, Google and Cursor's own Grok 4.6 on a single subscription.
  3. Cursor Router routes routine work off the frontier and keeps hard tasks on the strongest model, with the vendor publishing a 60% online A/B cost saving.
  4. Cloud agents keep working after the laptop closes, viewable from desktop, web and iOS.

What are Cursor's pros and cons?

What's great

  • Fluid, fast in-editor agent that most developers can adopt in minutes.
  • Model choice now covers Claude, GPT, Gemini and Cursor's own Grok 4.6 on the same subscription, with Cursor Router to dispatch routine work off the frontier.
  • Cloud agents, CLI, web app and iOS app ship on the same account, so a task can be assigned from anywhere and reviewed back in the editor.
  • Public ARR around $2 billion as of 2026-03 and a Gartner Leader placement in 2026-05 are independent anchors of adoption that match the vendor's own customer claims.

Watch-outs

  • Closed-source and subscription-priced. Heavy daily use pushes buyers from $20/month Pro up through $60 Pro+ and $200 Ultra to keep pace with agent limits.
  • The agent can over-edit on vague instructions. Diff review is the main guard, especially on large multi-file changes.
  • SpaceX closed its acquisition of Cursor's parent Anysphere on 2026-08-14 and the Cursor team has joined SpaceXAI. Model neutrality today is unchanged, but procurement teams that need a multi-vendor commitment on a 12-18 month horizon have one fewer clean answer.
  • No first-class headless or CI mode. Cursor is an interactive editor with a CLI and background agents, not a scriptable terminal agent in the same mould as Claude Code or Codex CLI.
  • Pro-tier fast-request limits can throttle a heavy day of agent work.
  • No public SLA on data retention or training use beyond the Privacy Mode opt-in, which matters for teams that cannot send code to a multi-tenant cloud service.

Cursor pricing

Free Hobby tier. Pro $20/month, Pro+ $60/month, Ultra $200/month billed monthly. Teams from $40/user/month, Enterprise custom. All paid tiers include access to Anthropic, OpenAI, Google and Cursor's own Grok 4.6.

  • Hobby

    Free

    • No credit card required
    • Limited Agent requests
    • Access to Composer
  • Pro

    $20 / month

    • Extended limits on Agent
    • Generous limits for Grok
    • Access to frontier models (Claude, GPT, Gemini, Grok 4.6)
    • MCPs, skills and hooks
    • Cloud agents
    • Bugbot on usage-based billing
  • Pro+

    $60 / month

    • 3x Pro limits on Agent
    • Generous limits for Grok
    • All Pro features
    • Access to Grok Bot
  • Ultra

    $200 / month

    • 20x Pro limits on Agent
    • Generous limits for Grok
    • All Pro+ features
    • Priority access to new features
    • Access to Grok Bot
  • Standard (Teams)

    $40 / user/mo

    • Everything in Individual
    • Centralised team billing and administration
    • Team marketplace for internal rules, skills and plugins
    • Agentic code reviews with Bugbot
    • Cloud agents and automations with shared team context
    • SAML/OIDC SSO, team-wide privacy mode
  • Premium (Teams)

    Custom (user/mo)

    • 5x Standard limits on Agent
    • All Standard features
    • Access to Grok Bot
  • Enterprise

    Custom

    • Everything on Teams
    • Pooled usage, invoice/PO billing
    • SCIM seat management
    • Repository, model and MCP access controls
    • Auto-run, browser and network controls
    • Audit logs, service accounts, AI code tracking API
    • Priority support and account management

See current pricing on cursor.com ↗Compare Cursor alternatives →

Frequently asked questions

Which models can Cursor use?
Anthropic Claude, OpenAI GPT, Google Gemini and Cursor's own Grok 4.6, all selectable per request. Cursor Router, generally available for Teams and Enterprise since 2026-07-22, classifies each request and routes it to the most capable model for the task.
Is Cursor free?
Yes. The Hobby tier is free with limited Agent requests and access to Composer. Paid individual tiers are Pro at $20/month, Pro+ at $60/month (3x Pro's Agent limits) and Ultra at $200/month (20x Pro's limits). Teams start at $40/user/month and Enterprise is custom. All paid tiers include access to Anthropic, OpenAI, Google and Cursor's own Grok 4.6.
Is Cursor open-source?
No. Cursor is a closed-source fork of VS Code. The Hobby tier is free to use, but the editor itself is proprietary. The closest open-source alternative in the directory is Cline, which is Apache-2.0 and bring-your-own-key.
Is Cursor owned by SpaceX now?
Yes. On 2026-08-14 Cursor's parent company Anysphere officially closed its acquisition by SpaceX and the Cursor team joined SpaceXAI. The deal was first announced in June 2026 at a reported $60 billion valuation.
What is Cursor Router?
An automated request router, generally available for Teams and Enterprise since 2026-07-22. It classifies each request and routes it to the most capable model for the task, with three modes (Intelligence, Balance, Cost) along the cost-intelligence frontier. Cursor's own online A/B testing reports roughly 60% cost savings versus routing every request to a frontier model.

Cursor alternatives

  • GitHub Copilot

    The safer default for teams standardised on GitHub, with more conservative agent-mode autonomy and the most mature org-level governance and IDE coverage. Cursor is a more polished but closed-source subscription whose parent is now SpaceX.

  • Cline

    A free, open-source (Apache-2.0), bring-your-own-key alternative with zero vendor lock-in, for developers who want full control over which model runs and are willing to trade Cursor's polish for unbounded usage-based API cost instead of a subscription.

  • Junie

    JetBrains' own IDE-embedded agent with Agentic Debugging (drives the real IDE debugger) and Plan Mode, for teams standardised on a JetBrains IDE rather than Cursor's VS Code fork.

  • Compare the whole category

    Every agent in Coding agents, side by side on the same fields.

Anything to add?

This page is researched and re-verified by our editors — but the people who actually use Cursor know things we cannot see from the outside.

Write a reviewUsed Cursor? Your experience helps the next buyer decide.
Get a quote from CursorTell us what you need and we make the introduction.
Get weekly Coding agents updates

New agents and fresh verdicts. One short email, every week. One-click unsubscribe.

Advertise

Reach buyers mid-decision. Put your brand in front of them — sponsor slots are open to any advertiser (see the content policy); listed agents can go Featured at the top of their category.

Featured — top of your category →Sponsor this page →

The evidence

Facts re-verified

How we checked this

The category adoption leader. Real cost is closed-source lock-in and the now-confirmed SpaceX ownership; the upside is the most complete in-editor agent surface and a model roster that now includes its own Grok 4.6.

What changed, and what we read (4 updates)
  1. Read cursor.com (docs), cursor.com (official), cursor.com (official), trustradius.com (review aggregator)

  2. API note

  3. API availabilityAlternativesConsFaq itemsHow it worksIntegrationsKey featuresLimitationsNot forPricingProsTaglineUse casesVerdictWhat isWhat it doesWho forWhy use

  4. Tier priority

Reviewed by The Agents Index Editorial, Research desk.

Advertise

Reach buyers mid-decision. Put your brand in front of them — sponsor slots are open to any advertiser (see the content policy); listed agents can go Featured at the top of their category.

Featured — top of your category →Sponsor this page →

Advertise

Reach buyers mid-decision. Put your brand in front of them — sponsor slots are open to any advertiser (see the content policy); listed agents can go Featured at the top of their category.

Featured — top of your category →Sponsor this page →

Your agent, listed

Built an AI agent or agentic tool? Get it researched, quality-gated and listed.

Submit it free