Cursor Composer 2.5 ships multi-agent orchestration — parallel sub-agents for refactor, test, doc generation in one IDE session

Cursor's Composer 2.5 update adds multi-agent orchestration: a planner agent decomposes a task into sub-tasks, then dispatches parallel sub-agents for refactor, test-writing, and documentation generation against the same code state. The update lands as a direct competitive response to Claude Code's terminal-native multi-agent workflows and Devin's cloud-agent pattern.

The architectural pattern matters because it brings the "swarm of specialized agents" design into the IDE-native tier. Until now, multi-agent orchestration lived in cloud-hosted products (Devin, OpenHands) or terminal-CLI tools (Claude Code). Cursor 2.5 collapses that split — the planner-and-dispatcher pattern is now the default in-IDE experience for paying users.

The competitive question is whether multi-agent IDE patterns generalize across model providers. Cursor's 2.5 release uses Claude and GPT family backends interchangeably; Anthropic's Claude Code is single-provider by design. The orchestration layer is becoming the lock-in surface, not the underlying model. See our analysis →.

Lushbinary — AI coding agents 2026 → · Artificial Analysis — coding agents comparison →