OpenAI Ultrafast speed tier for Codex and the API
According to OpenAI, Ultrafast offers up to 8× faster token generation in Codex and up to 6x in the API.
OpenAI's DevDay 2026 recap lists more than 20 announcements, including new tools for developers in Codex and the API. The announcement is dated September 29, 2026. The GPT-6.1 Sol launch and the Dots agents are covered in our GPT-6.1 Sol story and our Dots story.
Ultrafast
According to OpenAI, "Ultrafast is our premium speed tier for workloads where speed matters most. Ultrafast offers up to 8× faster token generation (300 tokens per second) in Codex and up to 6x in the API." These are OpenAI's figures, not our measurements.
OpenAI writes: "GPT‑6 Astra Ultrafast is available today in the API and in ChatGPT Work and Codex on Pro 500 and Enterprise plans. GPT‑6.1 Sol Ultrafast is coming soon." Our reading: for now the Ultrafast model available in the API is GPT-6 Astra; Sol Ultrafast has no date.
Pro 500
Per OpenAI, "the new Pro 500 plan offers our highest usage allowance at 25 times the ChatGPT Plus allowance and includes access to Ultrafast."
Decisions API
According to OpenAI, "Decisions API enables real-time decision-making by focusing Luna's intelligence on a specific set of user-defined questions with finite pre-defined answers." The page adds: "Available in limited preview today with a broad release planned in the coming days." Developers supply context using text or images, and get back answers they can use to classify content, route requests, or choose an agent’s next action.
Agents API with Computer use
OpenAI writes: "The Agents API now supports computer use so developers can build agents that interact with software to complete tasks. It also brings Codex's multi-agent capabilities, tool search, tool calling, and context compaction into your application." "Available through the API and in Codex and ChatGPT Work on Pro 500 and Enterprise."
Bedrock Managed Agents
The page says: "Now you can use Bedrock Managed Agents to build OpenAI agents that run entirely in AWS."
Private Intelligence
OpenAI writes: "OpenAI Private Intelligence helps businesses use frontier AI with greater confidence that their data is protected." It adds: "Our preview of Private Inference, coming this fall, combines confidential computing with strict, verifiable controls."
What we don't know yet
- Prices: the page gives no prices for Ultrafast or Pro 500.
- The date for GPT-6.1 Sol Ultrafast: the page says it is coming soon.
- The exact date of the broad Decisions API release: the page says "the coming days."
- What Luna is: the recap names it but does not describe it.
Not independently verified: every figure above is from OpenAI's announcement.
Source: OpenAI