OpenAI has officially delayed the broad launch of its next-tier model, GPT-5.6, at the direct request of the United States federal government. Instead of a general release, the model will initially launch as a limited preview restricted to a select group of corporate clients. This unprecedented move underscores growing friction between rapid AI development and national security protocols.
The federal government is maintaining a tight grip on this testing phase, with officials reportedly set to approve every individual access request during the preview period. According to reports from The Verge, the administration is particularly apprehensive about potential security vulnerabilities and the unpredictable capabilities of the new architecture.
- Individual vetting for all corporate users
- Strict federal oversight on model deployment
- Focus on mitigating unforeseen safety risks
