DeepSeek and Moonshot used OpenAI GPT-6 Astra outputs for model training

Chinese AI developers DeepSeek and Moonshot AI relied on outputs from OpenAI's GPT-6 Astra model to train their recent releases via distillation, according to a report citing people with direct knowledge. OpenAI had previously linked a related distillation campaign beginning in July to individuals associated with Moonshot AI. The practice remains a point of contention among leading labs as it enables faster capability replication at lower cost amid US-China technology competition.