Cursor Ships Composer 2.5: Custom Agentic Model Hits 79.8% SWE-bench Multilingual at 10-60x Lower Cost Than Frontier Models

Cursor Blog·high signal

Cursor released Composer 2.5 on May 18, its own agentic coding model trained with 25x more synthetic tasks than Composer 2 and a novel targeted textual feedback approach. It scores 79.8% on SWE-bench Multilingual (within 1 point of Opus 4.7, ahead of GPT-5.5) and ranks 3rd on the Artificial Analysis Coding Agent Index at $0.50/$2.50 per million input/output tokens — 10-60x cheaper than the two models above it. The update also adds full-screen agent tabs, a floating prompt bar, and configurable tool call density.