Cursor releases Composer 2.5 — matches Claude Opus 4.7 on coding at one-tenth the cost

Summary

Cursor shipped Composer 2.5 on May 18, its in-house long-horizon coding model built on Moonshot's open-source Kimi K2.5 checkpoint with 25× the synthetic training task volume of Composer 2. The model scores 79.8% on SWE-Bench Multilingual and 63.2% on CursorBench v3.1 — matching Claude Opus 4.7 and GPT-5.5 on key coding benchmarks. Standard-tier pricing is $0.50/M input and $2.50/M output tokens (roughly one-tenth Claude Opus pricing); fast tier is $3.00/$15.00. Cursor included double usage for