OpenAI expands ultrafast API access as new model features emerge

On September 27, 2026, X platform user @imjustnewatai reported that OpenAI has begun updating its platform and API documentation with Ultrafast API details, suggesting the company is preparing to expand access to the feature. The Ultrafast API, which allows for processing speeds up to 14 times faster than the Standard mode, is currently available only to select customers. It is powered by GPT-5.6 Sol and can handle up to 750 tokens per second. OpenAI is also working on a new speed option for the Responses API Playground, offering users Standard, Fast, and Ultrafast choices, though the feature remains hidden. The company is set to unveil more details at its DevDay event on September 29. Additionally, OpenAI has launched the GPT-6 series, including GPT-6 Sol and GPT-6 Astra, with uncertainty surrounding whether all models will support Ultrafast capabilities. The expansion of Ultrafast API access is expected to follow the DevDay event, where OpenAI will likely provide further insights into its new models and features. While the exact timeline for broader access remains unclear, the company's recent actions indicate a strategic push to enhance its API offerings and meet growing demand for faster processing speeds.