Why is OpenAI lobbying against Chinese open models after signing a pro-open letter

the trump administration prefers targeted national security bans on specific chinese open-weight models over broad restrictions, the new york times reports. models like kimi k3 currently trail western rivals on cybersecurity benchmarks but still threaten market dominance. openai and anthropic privately lobby the white house to restrict these models to escape severe price pressure, even though openai publicly signed a petition opposing open-weight regulation. this competitive maneuvering forces developers relying on affordable chinese apis to prepare for sudden access disruptions.