Anthropic names Zhipu for the first time, accusing GLM-5.2 of distilling Claude and OpenAI models

robot
Abstract generation in progress
According to Beating monitoring, Anthropic’s national security policy lead Tarun Chhabra first named Zhipu at the Aspen Security Forum. He accused GLM-5.2 of distilling Claude and OpenAI models.

Distillation is training another model using answers generated by a stronger model. The technology itself is quite common; the controversy is whether it bypasses service terms and massively scrapes competitors’ output.

Chhabra said that nearly all leading Chinese AI labs are distilling US models. Anthropic is currently banning millions of related accounts every week. US frontier models are currently ahead of China by about 6 to 9 months. He believes that without this kind of distillation, the gap could have been as large as 12 to 18 months.

Previously, Anthropic had also named DeepSeek, Moonshot, MiniMax, and Alibaba.
ZHIPU AI-9.31%
View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • Comment
  • Repost
  • Share
Comment
Add a comment
Add a comment
No comments
  • Pinned