OpenAI adds Ultrafast mode for GPT-6 Astra
OpenAI added Ultrafast mode for GPT-6 Astra in the Responses API to lower output token latency.
Models:gpt-6-astra
What changed
OpenAI added Ultrafast mode for GPT-6 Astra in the Responses API. Developers can use gpt-6-astra with service_tier: “ultrafast” to reduce the time between generated output tokens. The mode is available to API customers, subject to rate limits, with global processing and US data residency. EU and other regional inference residency are not supported.
What it means for Zabio customers
GPT-6 Astra is available in the Zabio catalog in the frontier tier and is used by auto routing. Developers calling this model through Zabio can use its standard capabilities, while regional inference residency remains limited to global processing and US residency.
What Zabio does
Nothing changes for Zabio customers.
Source
Added Ultrafast mode for GPT-6 Astra in the Responses API. Use gpt-6-astra with service_tier: “ultrafast” to reduce the time between generated output tokens.
Original announcement: OpenAI API changelog
Published automatically from the official announcement, minutes after it appeared. Reviewed by a person within the day.
Source: OpenAI API changelog (2026-09-29)