Cursor launches agent mode — direct competitor to Claude Code and Codex

Cursor, one of the fastest-growing startups in the history of AI developer tools, has announced the launch of the next generation of its product. It is no longer just an editor with smart autocomplete — Cursor is transitioning into agent mode, capable of independently solving developer tasks from setup to delivery. In parallel, the startup finds itself in direct competition with two of the world's largest AI laboratories — Anthropic and OpenAI.

Anysphere, the company behind Cursor, has followed a path typical of the best AI startups: it took a well-known tool (VS Code), added deep integration with language models, and created a product that developers loved for its speed and convenience. By early 2025, Cursor had become one of the most discussed IDEs among professional programmers, and the company's ARR exceeded $100 million — a record pace in the history of the category. Until now, Cursor had been built on models from third-party providers — Claude from Anthropic, GPT-4 from OpenAI, and others.

The startup was essentially a distributor of computing power with a higher-quality interface: deep codebase context, smart chat, built-in commands for refactoring. But now, with the launch of agent mode, Cursor is taking a step toward autonomous task execution — and inevitably comes face-to-face with the very companies whose APIs it uses. What does agent mode look like in practice?

Previously, a developer would give Cursor a specific task: "write a function," "fix this bug." Now the tool can accept a high-level task — for example, "add OAuth2 support to this project" — and independently go through the entire process: study the code structure, identify the necessary files, make changes, run tests, and propose a final result. This is a fundamentally different level of autonomy.

Competition in this field has already taken shape. Claude Code from Anthropic — a terminal agent that works with the file system, runs commands, and iterates on tasks. Its main strength is deep integration with Claude and support for the MCP protocol.

Codex from OpenAI — a cloud agent that executes tasks asynchronously in isolated containers, capable of working on multiple tasks in parallel. Both products were released or significantly updated in 2025, and both are targeting the same audience: developers who need not an assistant for hints, but a full-fledged AI colleague. For Cursor, this is a strategic crossroads.

Agent mode is a logical development of a product that developers already love. But Cursor depends on models from Anthropic and OpenAI. By launching an agent, the startup is changing its position in the ecosystem and risks aggravating relations with suppliers who have now become direct competitors.

The market for AI developer tools is undergoing rapid consolidation. A few months ago, there was room for dozens of players: GitHub Copilot, Cursor, Windsurf, Replit, Devin, Codeium. Today, the largest laboratories are building their own agent products and competing not for a single feature, but for the right to become an infrastructure layer in a programmer's daily work.

For the average developer, this means one thing: in the coming months, the quality of AI agents will grow sharply, and competition will force all parties to quickly improve their tools. Who will emerge as the winner from this race is still unclear. But developers will only benefit from this struggle.