OpenAI’s GPT-5.6 Launches After U.S. Government Delay

OpenAI’s GPT-5.6 models are launching Thursday after an earlier restriction limited access to select partners under U.S. government pressure. The Department of Commerce approved the public launch after the Center for AI Standards and Innovation ran additional tests, according to the article. OpenAI criticized the hold, arguing that it kept top tools away from developers and companies.

The launch approval does not settle the broader process for future frontier AI releases. The article notes that binding standards for releasing such models still do not exist, even as government review expectations are becoming more visible. For AI teams, the practical takeaway is to watch not only model capability, but also the approval environment that can affect access timelines.

OpenAI says GPT-5.6 Sol beats Anthropic’s Claude Mythos 5 on several benchmarks. On TerminalBench 2.1, Sol scored 88.8 percent, Sol Ultra reached 91.9 percent, and Mythos 5 scored 88 percent. The article also says Sol matched Mythos 5 on cybersecurity tasks while using only a third of the tokens, and lists Sol pricing at $5 per million input tokens and $30 per million output tokens.

Developers and companies evaluating GPT-5.6 should compare benchmark gains against real workload costs, especially token usage. The reported pricing and efficiency claims make cost-per-task a key metric, not just headline scores. The unresolved standards question also means future model access could depend on testing and government review processes that are still taking shape.