Microsoft and Nanjing Univ. Introduce LoopsBench for Long-Horizon Agent Coding
jiqizhixin · x · 2026-08-27
Microsoft and Nanjing University introduced LoopsBench, a benchmark for evaluating the long-term software engineering capabilities of AI agents. It includes 112 tasks, over 5,300 dev units, and 8 languages with a median dependency depth of 6. It tests how agents maintain plans, code, and test states over extended execution, not just single-task resolution. Results show even the best config achieves only a 25% task resolve rate, though continuation mechanisms improve performance.
More from coding & agent
- Musk amplifies dev's verdict: Grok Bot easier to use than Codex and Claude — elonmusk · 2026-10-15
- Blob0 launches: a macOS file-manager-style desktop client for S3 and Cloudflare R2 — yihui_indie · 2026-10-15
- The 6-Layer AI Engineer Skill Stack for 2026: Vectors to Multi-Agent Ops — mdancho84 · 2026-10-15
- Building a Mechanical Engineering Agent Workflow in ChatGPT: 30-Node Canvas Shows Promise — burhop · 2026-10-15
- Open-source Harness Engineering course hits 19k stars, shows $9 vs $200 gap between failing and working agent setups — thisguyknowsai · 2026-10-15
- A 6-person team shipped healthcare prior-auth AI agents in 3 months — here's the playbook — alex_verem · 2026-10-15