OpenAI's GPT-5.6 models are set to launch on Thursday, following a delay imposed by the U.S. government. The models were initially restricted to select partners in late June due to pressure from federal authorities. The Department of Commerce approved the public launch after the Center for AI Standards and Innovation conducted additional tests, according to Axios. OpenAI criticized the delay, stating it prevented developers and companies from accessing the best tools. The company emphasized the need for binding standards for model releases, as outlined in Trump's latest AI executive order, which remain unimplemented.
GPT-5.6 Sol Ultra led in the TerminalBench 2.1 coding benchmark with a score of 91.9 percent. It outperformed Anthropic's Claude Mythos 5, which scored 88.0 percent, and Google's Gemini 3.1 Pro Preview, which scored 70.7 percent. On cybersecurity tasks, Sol matched Mythos 5's performance but used only a third of the tokens. OpenAI noted that Sol costs $5 per million input tokens and $30 per million output tokens, significantly lower than Anthropic's Fable 5, which runs nearly double at $10 per million input and $50 per million output.
OpenAI said the delay caused by the U.S. government kept the best tools away from developers and companies. The company criticized the lack of binding standards for releasing large language models, which were mandated by Trump's latest AI executive order. The approval process for the public launch of GPT-5.6 highlights the ongoing regulatory challenges in the AI industry.
Source: thedecoder
