AI RESEARCH
KAT-Coder-V2 Technical Report
arXiv CS.LG
•
ArXi:2603.27703v1 Announce Type: cross We present KAT-Coder-V2, an agentic coding model developed by the KwaiKAT team at Kuaishou. KAT-Coder-V2 adopts a "Specialize-then-Unify" paradigm that decomposes agentic coding into five expert domains - SWE, WebCoding, Terminal, WebSearch, and General - each undergoing independent supervised fine-tuning and reinforcement learning, before being consolidated into a single model via on-policy distillation. We develop KwaiEn, a modular infrastructure sustaining tens of thousands of concurrent sandbox instances, and scale RL.