List your company

Role Clarity on SWE hard (dev)

90.76Role Clarity ScoreChatDev
48.494459.467270.4481.4128Apr 3, 2026

Evaluation results

MethodLinks
ChatDevCEO=FT, CPO=FT, Backbone=Qwen, Role clarity regularization=Enabled2026.0490.760.3781
ChatDevCEO=FT, CPO=FT, Backbone=Llama, Role clarity regularization=with2026.0485.350.3523
ChatDevCEO=FT, CPO=Base, Backbone=Qwen, Role clarity regularization=Enabled2026.0464.40.1145
ChatDevCEO=Base, CPO=FT, Backbone=Qwen, Role clarity regularization=Enabled2026.0460.440.0749
ChatDevCEO=Base, CPO=FT, Backbone=Llama, Role clarity regularization=with2026.0456.550.0643
ChatDevCEO=FT, CPO=Base, Backbone=Llama, Role clarity regularization=with2026.0454.960.0484
ChatDevCEO=Base, CPO=Base, Backbone=Qwen, Role clarity regularization=Enabled2026.0452.95—
ChatDevCEO=Base, CPO=Base, Backbone=Llama, Role clarity regularization=with2026.0450.12—