Feb 23, 2026
Anthropic alleges that DeepSeek, Moonshot AI and MiniMax used approximately 24,000 fraudulent accounts to generate over 16 million exchanges with Claude for training their own models.
Shows competition over reasoning, coding and agent capabilities extending to access controls and training data. Distillation is a legitimate training method; Anthropic alleges unauthorized access and use in these campaigns. The date marks the disclosure, not when the campaigns began.