Cursor 2.0 Launches with Composer Model and Multi-Agent Architecture
Cursor 2.0 introduces Composer, a purpose-built coding model 4x faster than comparable AI, plus parallel multi-agent execution and a built-in browser.
Cursor’s Biggest Update Yet
Cursor 2.0, released on October 29, 2025, represents the most significant update to the AI-native code editor since its launch. The release introduces Composer — Cursor’s own purpose-built coding model — along with a completely redesigned agent-centric interface and the ability to run up to 8 AI agents in parallel.
The update cements Cursor’s position as the leading AI-native IDE, arriving just weeks before the company’s $2.3 billion Series D at a $29.3 billion valuation.
Composer: A Frontier Coding Model
Composer is Cursor’s proprietary language model, built specifically for software engineering tasks. According to Cursor, it runs approximately 4x faster than models of comparable intelligence, with most conversational turns completing in under 30 seconds.
The model powers a new agent-centric interface where agents, plans, and runs are first-class objects in the sidebar. Developers can manage multiple simultaneous agents, each with their own context and execution state.
Multi-Agent Parallel Execution
The headline productivity feature: run up to 8 agents simultaneously, each operating in isolated copies of the codebase via git worktrees or remote machines. This means you can have one agent refactoring authentication, another writing tests, and a third implementing a new feature — all at the same time.
Each agent maintains its own branch and execution context, preventing conflicts and enabling true parallel development workflows.
Built-in Browser
Cursor 2.0 includes a built-in browser that can pass selected elements and DOM context back to the agent. This is particularly useful for UI development — select an element in the browser, describe what you want changed, and the agent modifies the code accordingly.
What This Means
Cursor 2.0 shifts the paradigm from AI-assisted coding to AI-first development. The multi-agent architecture, combined with a purpose-built model, positions Cursor as not just a better code editor, but a fundamentally different way to build software — where developers manage AI agents rather than writing every line themselves.