OpenAI Plans Public Release of Its Strongest Model, GPT-5.6

OpenAI is preparing to release its newest and most powerful model, GPT-5.6, for public use following a period of delay linked to security evaluations conducted by US authorities. This move marks a critical phase in the competition over large AI models, where the focus extends beyond AI capabilities to include the manner of release and access rights.

Previously, access to GPT-5.6 was limited to a select group of trusted partners. According to a Reuters report, OpenAI will launch GPT-5.6 under the name Sol, accompanied by two less expensive variants called Terra and Luna. The delay in the rollout reportedly resulted from requests by the US government, which expressed concerns about the potential misuse of advanced models in cyberattacks or activities related to national security. These concerns arise amid intensifying competition between the United States and China over the most advanced AI technologies.

The release of GPT-5.6 coincides with broader efforts in Washington to impose prior reviews of advanced AI models before they become publicly available. Reuters also noted that the company Anthropic had restricted access to some of its most powerful models following export control orders, later reinstating certain services under enhanced protective conditions. GPT-5.6 enters the market at a time when governments are treating advanced artificial intelligence as a strategic asset with sensitivity comparable to that of semiconductor chips and cloud infrastructure.