OpenAI has introduced Ultrafast, a preview mode intended to make its GPT 5.6 Sol model respond and complete work far more quickly. The company says it could help businesses use a powerful general-purpose model in time-sensitive tasks without switching to smaller or specialised systems.

OpenAI claims Ultrafast can run at up to 14 times standard processing speed and generate as many as 750 output tokens per second. It is powered by OpenAI’s partnership with chipmaker Cerebras and is initially available only to a small group of customers, with wider access planned as capacity increases; proposed uses include incident response, customer support, market analysis and e-commerce.