Anthropic sent a letter to the U.S. Senate Committee on Banking, Housing, and Urban Affairs accusing Alibaba and operators associated with Qwen of large-scale extraction of Claude’s capabilities. Anthropic alleged that the operators used approximately 25,000 fraudulent accounts to extract capabilities from the Claude model, describing it as the 'largest known' model distillation attack. The incident occurred between April 22 and June 5, 2026, involving more than 28.8 million interactions with Claude. Model distillation uses outputs from a powerful model to train another model, potentially circumventing export restrictions on model weights. The letter, dated June 10, was addressed to Chairman Tim Scott and Ranking Member Elizabeth Warren. Anthropic believes the operators are affiliated with Alibaba and Alibaba Qwen, aiming to accelerate China’s acquisition of Anthropic’s advanced model capabilities. The concern centers on potential leakage of capabilities in software engineering, automated task execution, and agent reasoning. Alibaba has not responded to the distillation allegations.