OpenAI launches 'Ultrafast' mode, speeding up GPT-5.6 Sol by 14 times

OpenAI has unveiled a new operating mode called Ultrafast, designed to significantly speed up the performance of its most powerful model, GPT-5.6 Sol. According to a company blog post published Thursday, the new mode allows the model to work up to 14 times faster than standard processing, reaching as much as 750 output tokens per second. Output tokens refer to the individual pieces of text an LLM generates while interacting with a user.

OpenAI states that achieving real-time speed previously required choosing a smaller or more specialized model. Ultrafast, the company says, represents a new approach — delivering more useful work per second without sacrificing model capability. The feature is powered by OpenAI's partnership with chipmaker Cerebras.

Competition and use cases

Other AI companies, including Anthropic, have also released accelerated versions of their models — Anthropic offers a fast mode for Claude, though it does not match the speed OpenAI is now offering.

OpenAI suggests Ultrafast could be useful across several corporate workflows, including incident response, customer service and support, financial market analysis, and e-commerce.

Ultrafast is currently being released in preview and is available only to a small group of customers. OpenAI says it plans to expand access as capacity grows.