US clears broad rollout of GPT-5.6 after month of restrictions

The U.S. Department of Commerce has approved the broad rollout of OpenAI's GPT-5.6 model, ending a month of restrictions imposed at the request of the U.S. government. Axios first reported the news, citing a person familiar with the matter.

A month ago, OpenAI held back the full public launch of GPT-5.6 at the explicit request of the U.S. administration. Instead of a broad release, the company limited access to a narrow group of partners whose data it shared with the government. That move was part of the previously reported restriction on the Sol variant's availability, introduced at Washington's request, as nowosci.ai covered earlier.

What the government reviewed

Testing was conducted by the Center for AI Standards and Innovation, a Department of Commerce unit responsible for assessing the safety of the latest AI models before their public release. OpenAI sent a technical team to Washington to answer officials' questions throughout the verification process.

The entire procedure rests on a voluntary mechanism established by a Trump administration executive order, which requires developers of the most powerful AI models to report them to the government for up to 30 days before they reach even trusted business partners. GPT-5.6 was one of the first models to go through this process at such scale.

Three variants of one model

GPT-5.6 is not a single model but a family of three variants with different levels of capability and cost. Sol is, according to OpenAI, the most advanced version, built for the most demanding use cases. Terra is a mid-tier option, while Luna is the cheapest variant, optimized for lower running costs.

This distinction matters in practice for companies planning deployments: the government's decision covers the broad release of the entire model family, not just the flagship Sol variant, which had previously gone only to a narrow group of partners vetted by the administration.

What it means for the market

The month-long delay in GPT-5.6's launch showed just how directly the U.S. administration can influence the commercial timelines of leading AI labs. For OpenAI, it meant losing momentum against rivals such as Anthropic and Google DeepMind, which kept rolling out their own releases without comparable restrictions.

The approval for broad deployment signals that the review process has concluded without any additional conditions disclosed publicly. For companies using OpenAI's API, it means full access to GPT-5.6 starting Thursday, with no need to apply for vetted-partner status.

For Polish companies and developers using OpenAI's models, the key question will be how quickly GPT-5.6 becomes available outside the U.S. market. The restrictions so far mainly applied to access within the United States, but OpenAI's global rollouts typically follow shortly after.

The precedent set in this case could also shape how future frontier-model launches from other companies play out. Anthropic, Google, and xAI are subject to the same voluntary reporting mechanism stemming from the Trump administration's executive order.

Sources: OpenAI gets US approval for broad GPT-5.6 rollout, Axios reports (investing.com), OpenAI gets US approval for broad GPT-5.6 rollout, Axios reports (finance.yahoo.com)