Chinese AI agent outperforms Anthropic’s Claude Code in autonomous research

A Chinese artificial intelligence (AI) system has topped an international ranking for autonomous scientific research, pulling ahead of Anthropic’s Claude Code and other top agents. As of Tuesday, the Zhejiang University-led Qiushi Engine held the top overall spot on the ResearchClawBench leaderboard, with Open Science Desktop in second and Claude Code in third. ResearchClawBench tests the ability of AI agents to independently carry out research and compares their results against reference papers