US government calls on OpenAI to restrict GPT-5.6

According to information obtained by CNN, the United States administration has requested that OpenAI restrict the distribution of its GPT-5.6 artificial intelligence model, granting access only to certain government-approved organizations. It is stated that this request stems from the model's powerful capabilities in the field of cybersecurity.

Sources close to OpenAI reported that the Washington administration considers GPT-5.6 to be at the same risk level as Anthropic's Mythos model, which was subjected to export controls in recent weeks. It is reported that OpenAI has agreed to release the model in a limited capacity.

At the same time, the US-based website The Information obtained a memo sent by OpenAI and its CEO Sam Altman to employees. The note stated, "We have made it clear to the US government that this is not the model we prefer in the long term. We will work with them and others in the industry to develop a more sustainable approach for future versions." Additionally, it was noted that the government is approving access to the model on a per-customer basis.

White House officials stated that they are continuing to work with leading artificial intelligence research laboratories to manage the risks that the technology may pose. OpenAI has not yet made a statement regarding the matter.

On the other hand, on June 13, the administration of President Donald Trump implemented a similar export control decision for Anthropic's most advanced artificial intelligence models, "Mythos" and "Fable." Citing security concerns, access to the models by foreign governments, companies, and individuals was prohibited, resulting in Anthropic closing access to its models for global users.