Chinese labs reportedly distill OpenAI GPT-6 Astra outputs
Original title: OpenAI's Astra becomes the main target for Chinese AI labs after Anthropic tightens its defenses
The Information says DeepSeek and Moonshot used Astra, including hidden reasoning, to train Kimi K3.1 Pro and V4.1-Flash 1017.
According to The Information, Chinese AI labs have shifted distillation efforts from Anthropic’s Claude to OpenAI’s GPT-6 Astra. People familiar with the matter say DeepSeek and Moonshot AI relied heavily on Astra outputs for newer models, which requires coaxing models to reveal hidden reasoning that both vendors prohibit. Anthropic says newer defenses cut extractable data by 90 to 99 percent. In late September, OpenAI reported a coordinated distillation campaign it linked to people associated with Moonshot. Moonshot declined to comment and DeepSeek did not respond.
Unconfirmed: The distillation claims come from The Information and unnamed people; DeepSeek and Moonshot have not confirmed them.
- Named trainee models are Kimi K3.1 Pro and V4.1-Flash 1017.
- Anthropic claims its new methods cut data attackers can pull by 90 to 99 percent.
- OpenAI says attackers changed tactics after each new defense.
- An OpenAI spokesperson said the company takes distillation very seriously.
- Organizations
- DeepSeekMoonshot AI
- Countries
- China
From the original article
1 min read at the-decoder.com
Chinese AI labs are now mainly tapping OpenAI's most powerful model, GPT-6 Astra, instead of Anthropic's Claude. The article OpenAI's Astra becomes the main target for Chinese AI labs after Anthropic tightens its defenses appeared first on The Decoder.
By Maximilian Schreiner · the-decoder.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.