OpenAI Publicly Releases 3 GPT-5.6 Models as U.S.-Requested Limits End

3 articles · Updated · CNBC · Jul 8

Thursday marks OpenAI’s public launch of GPT-5.6 Sol, Terra and Luna, about two weeks after the company restricted access to a small group of trusted partners tied to a U.S. government review.
OpenAI said it is ending that limited rollout because it favors broad access, while still working with Washington on a repeatable process for evaluating frontier models before full release.
GPT-5.6 Sol is OpenAI’s strongest model yet, the company said, with improved capabilities in coding, biology and cybersecurity; preview access is also being expanded globally.
The release follows Commerce Department clearance and mirrors rival Anthropic’s restored access after an export-control directive was lifted late last month.
The shift fits a June Trump executive order that gave agencies 60 days to build a voluntary federal assessment process for cutting-edge AI systems.

Sources

Left100%

CNBC53m ago

OpenAI Publicly Releases GPT-5.6 AI Models, Ends Government-Requested Limits

CNBC4h ago

U.S. Commerce Department Clears OpenAI for Broad GPT-5.6 Model Release

Storyboard184h ago

OpenAI set for wider GPT-5.6 launch after US approval: Report - Storyboard18

11 Sources

Can America's cautious AI strategy win against China's rapid, low-cost model deployment in the global tech race?

As the AI chip market splits, must global firms choose between American innovation and Chinese affordability?

GPT-5.6 Public Release: Capabilities, Government Oversight, and the New Era of AI Regulation (July 2026)

Overview

OpenAI's GPT-5.6 model series was broadly released to the public on July 8, 2026, after an initial restricted rollout to about 20 trusted partners. This staggered release followed a new U.S. government executive order, signed by President Trump, which required advanced AI models to be submitted for federal review before public launch. OpenAI complied by sending experts to Washington D.C. and allowing additional government testing. This process, designed to address national security and cybersecurity concerns, marks a significant shift in how advanced AI models are reviewed and released, highlighting growing government involvement in AI governance.

...