OpenAI has officially announced that it will publicly release its latest generation GPT-5.6 series of models this coming Thursday. Previously, the Trump administration, citing cybersecurity concerns, had requested OpenAI to restrict the rollout of this new model.

Media outlet Axios reported concurrently on Tuesday that the US government has now given the green light, permitting OpenAI to widely launch the three sub-models of GPT-5.6: Sol, Terra, and Luna.

The US government's intervention to temporarily halt the full release of the AI model has generated significant confusion and concern within the artificial intelligence industry. Currently, major AI companies and US regulators are jointly developing a more standardized framework for controlling the release of large models. A key control point is the model's ability to identify software cybersecurity vulnerabilities, aiming to prevent related functionalities from posing security threats to US domestic enterprises.

Both Anthropic and OpenAI have already weakened, or plan to weaken, their models' capabilities related to uncovering cybersecurity vulnerabilities when making them available to general users.