Cursor Launches a New AI Agent Experience to Take On Claude Code and Codex
.jpg)
Why it matters: Cursor's new product aims to regain market share from leading AI labs in the competitive developer and enterprise customer segments.
- Cursor launched Cursor 3, an "agent-first" coding product, to allow users to spin up AI coding agents for task completion.
- Jonas Nelle, Cursor's head of engineering, notes that the profession has "completely changed," making Cursor's original product less relevant going forward.
- Cursor 3 uniquely integrates an agent-first product with Cursor’s AI-powered development environment, allowing developers to prompt an agent in the cloud and review generated code locally, unlike desktop apps for Claude Code and Codex.
- OpenAI and Anthropic have launched their own agentic coding products, offering highly subsidized subscriptions that pressure Cursor's business model.
Cursor has launched Cursor 3, an "agent-first" AI coding product designed to compete with established tools like Anthropic's Claude Code and OpenAI's Codex. This new interface allows developers to offload entire tasks to AI agents, marking a significant shift from Cursor's original IDE-centric approach and integrating agent capabilities directly into its existing desktop app.
