In February 2026, Anthropic accused DeepSeek, Moonshot AI, and MiniMax of creating approximately 24,000 fraudulent accounts and sending over 16 million prompts to its Claude model