OpenAI's GPT-6 Astra Becomes Top Distillation Target for Chinese AI Labs

Chinese AI labs DeepSeek and Moonshot AI relied heavily on outputs from OpenAI's GPT-6 Astra when training their latest models, according to people familiar with the matter. Distillation remains a major driver of progress for Chinese open-weight models despite criticism from OpenAI and Anthropic.

• DeepSeek and Moonshot AI trained new models on GPT-6 Astra outputs
• Astra launched in early September and saw strong demand
• Distillation remains a key driver of China's open-weight model progress