Anthropic will launch a new safety system later this year. The update allows enterprise customers to retain data on their own cloud infrastructure.
The policy applies to the company's most capable models, such as Mythos and Fable 5. It addresses corporate demands for greater data privacy and control.
Anthropic still requires companies to keep data for 30 days for safety and misuse monitoring. This move contrasts with OpenAI, which is developing monitoring systems that do not require data retention.
The new option provides security assurances for corporate clients handling sensitive information.