OpenAI says Moonshot AI tried copying Kimi K3 thinking model
Technology
OpenAI says Chinese lab Moonshot AI tried to copy how its AI thinks, linking these attempts to Moonshot's Kimi K3 model launched on July 16, 2026.
According to OpenAI, there were efforts to mimic its tech without hacking or stealing user data directly.
OpenAI tightened security after 16,000 attempts
OpenAI noticed a surge of 16,000 extraction attempts in late July and quickly tightened security and teamed up with partners to stop it.
This situation highlights growing worries about AI tech being copied or misused.
So far, Moonshot AI hasn't responded to the claims.