OpenAI delays GPT-5.6 after Trump admin requests staggered release

OpenAI is delaying the full release of its next flagship model, GPT-5.6, following a request from the Trump administration over security concerns. CEO Sam Altman announced in a company Q&A Wednesday that the model will instead launch in a limited preview form restricted to a small set of enterprise customers. During this preview period, the government itself will approve each customer on a case-by-case basis. This marks a significant shift from the administration's earlier promise of a “speed wins” approach to AI development and exports.

Industry observers are drawing comparisons to the far stricter treatment of Anthropic. Earlier this month, the administration gave Anthropic an ultimatum to suspend access to its Mythos 5 and Fable 5 models, backed by an export control directive barring “foreign nationals”—including some of Anthropic's own non-U.S. employees—from accessing the technology. The uneven regulatory hand has sparked concerns that the promised pro-innovation stance is giving way to heavy-handed, company-specific controls that could stifle competition and create compliance chaos for AI vendors and their enterprise customers.