OpenAI is starting to preview Ultrafast mode for GPT-5.6 Sol. Powered by Cerebras, it hits up to 750 tokens per second — as much as 14 times faster than Standard. For now, it's only available to a small number of API customers.
OpenAI has tested Ultrafast in troubleshooting, research, customer service, financial analysis, and agent development. For agents that need to call a model repeatedly, faster single responses directly shorten the time it takes to complete the entire task.