News

US Didn’t Approve GPT-5.6, White House Clarifies Role in Frontier AI Launches

US Denies Approving OpenAI GPT-5.6 Release as White House Clarifies No Formal Clearance Was Required Despite Earlier Rollout Delay Over National Security and Frontier AI Safety Concerns

Written By : Poulami Saha
Reviewed By : Achu Krishnan

The White House has denied reports of formally approving OpenAI's GPT-5.6 for public release, saying no such approval process exists for commercial AI models. The clarification came amid media reports that the Trump administration had approved the company's newest cutting-edge technology following weeks of discussions regarding national security concerns.

According to a White House official, the administration did not ‘approve’ GPT-5.6 since the U.S. government does not have any existing legislation that compels companies to seek permission prior to deploying new AI technologies.

US Clarifies it Never Approved GPT-5.6

OpenAI had earlier announced the rollout via its official X handle. The new model will be available in three versions, namely the Sol, Terra, and Luna. The confusion stems from OpenAI's decision to postpone GPT-5.6's public launch in June after the US government requested a staggered rollout. 

However, officials expressed concern that frontier AI systems with advanced capabilities in cybersecurity, biology, and coding could be misused if released without additional safety evaluations.

During the postponement period, OpenAI restricted access to select partners and continued testing and talks with relevant federal bodies. It should be noted that OpenAI had no objection to the release of its product in a phased manner but accepted the government's demands.

AI Governance Debate Intensifies

The initial reports claimed that the broader release was approved by the government. The White House clarified that its engagements with OpenAI had been purely advisory in nature and not regulatory in any way.

The release includes the flagship GPT-5.6 Sol model alongside lower-cost Terra and Luna variants, which target different user segments and workloads. The company says the models deliver stronger performance in software development, scientific reasoning, and cybersecurity tasks.

The episode reflects the growing influence of governments over frontier AI development without introducing formal licensing requirements.  It becomes clear to the AI industry that governments might not be approving AI launches, but are actively influencing when and how the most powerful models will become available.

Also Read: OpenAI Unveils Codex Micro, Marking its First Step into Consumer Hardware

UAE Proptech Keyper Raises $11M Series A to Scale Digital Rent Payments

Call of Duty Classics Return: Black Ops and Black Ops 2 Now Native on PS5 and PS4

AI Ads Get a Transparency Upgrade as Google Rolls Out New Disclosure Labels

Craig Burnatowski Denies Abuse Allegations After Fantasy & Magic Con Drops Him from Guest Lineup

Meta's Iris AI Chip Enters Production as Company Scales AI Infrastructure