Hobby
Free
Limited features
Anysphere's AI-native development platform. Cursor 3 (April 2026) rebuilt the editor around an Agents Window: a unified workspace for managing local, cloud, and background agents across multiple repositories. Powered by Composer 2 (Anysphere's in-house model), Claude Opus 4.7, GPT-5.5, and Gemini 3 Pro, with JetBrains ACP integration and an Agent CLI.
+Pros
- Most popular AI-native IDE — large community, extensive third-party extensions, and broad model support
- Background Agent runs autonomously in cloud VMs: clones repos, edits files on a branch, and opens PRs
- BugBot automatically reviews PRs on GitHub and leaves actionable comments linked to Cursor
- Plan Mode for complex multi-step reasoning before execution, with Mermaid diagram output
- Multi-agent support: parallel sub-agents for concurrent task execution
- Team Plugin Marketplace and MCP Apps for one-click tool integrations
- JetBrains IDE support since March 2026 — not limited to VS Code fork
−Cons
- Proprietary fork of VS Code — not open source, and deviates from upstream VS Code over time
- Background Agent and cloud VMs are usage-billed on top of the subscription — costs can escalate
- BugBot requires Max mode (Pro plan minimum) — not available on the free Hobby tier
- Security concerns with background agents: auto-runs commands, prompt injection risk acknowledged in official docs
- Teams plan at $40/user/mo is significantly pricier than GitHub Copilot Business ($19/user/mo)
Pricing
Pro
$20/mo
Full features, Composer 2
Pro+
$60/mo
Higher limits, background agents
Ultra
$200/mo
Maximum usage
Teams
$40/user/mo
Team management
Enterprise
Custom
SSO, compliance
Cursor is the AI-native development platform from Anysphere. Originally a VS Code fork with AI woven into every part of the editor, it grew into the fastest-scaling SaaS company on record — around $2B ARR and one million paying customers by early 2026. Cursor 3 (April 2026) rebuilt the interface around an Agents Window: a single workspace for orchestrating local, cloud, and background agents across multiple repositories.
The product spans a continuous autonomy slider — from Supermaven-powered Tab autocomplete and Cmd-K inline edits, through Composer for multi-file edits, up to Background Agents that clone your repo to a cloud VM, run tests, and open PRs. Cursor 3 also introduced Composer 2, Anysphere's own in-house frontier coding model, alongside Claude Opus 4.7, GPT-5.5, and Gemini 3 Pro. JetBrains integration via the Agent Communication Protocol extended the surface beyond the standalone editor.
Key Features
• Agents Window (Cursor 3): unified workspace for local, cloud, and background agents
• Background Agents: up to 8 parallel cloud agents on isolated VMs, PR-ready
• Composer 2: Anysphere's in-house frontier coding model with high Pro limits
• Multi-model: Claude Opus 4.7, GPT-5.5, Gemini 3 Pro, plus BYOK
• Tab completion powered by Supermaven (acquired 2025)
• Multi-repo workspaces and plugin/MCP marketplace
• Automations: event-triggered agents (PR review, CI, alerts)
• JetBrains ACP integration: agent features in IntelliJ, PyCharm, WebStorm
• Agent CLI and Slack integration for headless workflows
Version History
Development environments for cloud agents: dedicated sandboxed VMs per agent run. Context Usage Breakdown introduced in v3.3 (May 2026). Multitask, Worktrees, and Multi-root Workspaces in v3.2 (Apr 2026). Cursor 3.0 (Apr 2026) launched a redesigned interface with the new Agents Window, tiled layout, and upgraded voice input.
Plugins system and Team Plugin Marketplace launched. MCP Apps for one-click MCP server installation. Sandbox Network Access Controls for background agents. Async subagents for parallel task execution. JetBrains IDE support released (Mar 2026).
1.0 milestone: BugBot for automated PR code review (GitHub integration), Background Agent GA (cloud VM agents that clone repos and open PRs), one-click MCP installation, Memories first look, Jupyter Notebooks support, and coding agents in Slack.
Signature Snippet
Cmd-K in a Next.js file: "Add input validation and error handling to this route." Cursor analyzes the function, checks related types and imports zod, and rewrites the handler with proper error boundaries. Hand off the same task to a Background Agent and it works on a separate branch, runs the tests, and returns a PR with a video of the agent session.More in this Space
Prime Agent (Prime Intellect)
Prime Agent is an open-source (MIT) self-improving coding and research harness by Prime Intellect, built on a Recursive Language Model that runs inside a persistent IPython REPL and a Continual Harness that refines its own state via /refine. It took the same frontier weights from 30.2% to 95.5% on ARC-AGI-3 RHAE Best@1, powered daemon-backed sessions, and ships no hosted offering.
OpenCodeReview (ocr)
OpenCodeReview (ocr) is Alibaba's open-source AI code-review CLI. It reads git diffs and drives an LLM agent through a hybrid architecture where deterministic pipelines handle file selection, rule matching, and comment positioning while the agent does the reasoning. Built on two years of internal use at Alibaba scale, it ships a multi-language ruleset and consumes roughly 1/9 of the tokens of general-purpose agents.
Delta (by Zed)
Multiplayer environment from Zed for coding with agents. Threads unify the agent conversation, the code edits, and human review in one shareable artifact, replacing the branch/PR loop. Built on DeltaDB, a CRDT-based version control layer that records every edit between commits alongside the prompts and reasoning that produced them.