Verified harness rankings
Verified rankings of AI coding-agent harnesses on real benchmarks.
Unofficial run on Arnesa Host·Terminal-Bench 4.0.0 (16/66 tasks)·Aider Polyglot 1.0.0 (30/225 tasks)
No benchmark runs for this model yet
No published harnesses have ranked runs on this model under the selected filters.