SWE-CoWork Benchmark: Top Coding Agents Complete Under 8% of Tasks
TokenWave's SWE-CoWork benchmark evaluates coding agents as digital employees across 101 real-repo tasks in 11 languages, and finds even the best model fully completes only 8 tasks, under an 8% delivery rate.
2026-11-17 ~ 2026-11-18 · 2 related posts
- SWE-CoWork: coding agents pass 63% of tests but complete fewer than 8% of tasks — Tokenwave-AI · 2026-11-17
- SWE-CoWork benchmark: best model fully passes only 8 of 101 handoff coding tasks — omarsar0 · 2026-11-18