According to The Information, Chinese AI labs DeepSeek and Moonshot AI now rely mainly on outputs from OpenAI's most powerful model, GPT-6 Astra, rather than Anthropic's Claude. The report says the outputs were used to train the labs' latest models, including Kimi K3.1 Pro and V4.1-Flash 1017. The report cites people familiar with the matter.
According to the report, this kind of distillation only works if the labs can get a model to reveal its hidden reasoning steps. OpenAI and Anthropic both ban that practice. Moonshot declined to comment. DeepSeek did not respond to a request for comment.
The report says Claude used to be the main target, but Anthropic has tightened its defenses. Anthropic says its new methods cut the amount of data attackers can extract by 90 to 99 percent.
The report also says that in late September, OpenAI reported a coordinated distillation campaign that involved people linked to Moonshot. An OpenAI spokesperson said the company takes the problem very seriously. The spokesperson said attackers had adjusted their tactics each time OpenAI rolled out a new defense.
Source: The Decoder