OpenAI rolls out Ultrafast mode, boosting GPT 5.6 Sol speed fourteenfold

OpenAI unveiled Ultrafast, a new operating mode for its flagship GPT 5.6 Sol model that it says runs at fourteen times the speed of standard processing. According to the company's blog, the mode can output as many as 750 distinct text tokens each second, dramatically increasing real-time capabilities without resorting to smaller models. Ultrafast is being powered through a collaboration with chip manufacturer Cerebras and is now available in a limited preview to a select group of clients.

OpenAI highlighted potential uses in areas like incident response, customer support, financial market analysis, and e-commerce. While rivals such as Anthropic have released accelerated options, OpenAI argues its speed advantage is unprecedented. The firm indicated that broader rollout will follow as computational capacity expands.