OpenAI Speeds Up GPT-5.6 Sol by 14x With Ultrafast Mode

OpenAI has unveiled a high-speed processing option called Ultrafast for its flagship GPT-5.6 Sol artificial intelligence model. Currently running in a limited preview, this new mode is designed to deliver up to 750 output tokens per second. According to the company, this represents a 14-fold increase in speed compared to standard processing times, allowing the model to generate text at unprecedented rates.

The hardware driving this massive performance leap is powered by OpenAI's partnership with chipmaker Cerebras. Because of the heavy computational demands, the preview is initially restricted to a select group of customers. OpenAI plans to expand access to more users as its infrastructure capacity grows.

This release targets a common trade-off in AI deployment, where developers historically had to choose smaller, less capable models to achieve low latency. OpenAI claims Ultrafast shifts this paradigm by offering "more useful work per second" without sacrificing the intelligence of its top-tier model. While competitors like Anthropic offer a fast mode for Claude, OpenAI's new offering aims to surpass those speeds significantly.

For enterprise practitioners, this speed boost opens up new possibilities for automation in time-sensitive environments. OpenAI suggests that the rapid throughput of GPT-5.6 Sol in Ultrafast mode is highly suited for demanding corporate applications. These include live customer support, rapid incident response, real-time financial market analysis, and e-commerce operations where sub-second response times are critical to user experience.