OpenAI's September 29 DevDay recap introduced Astra Ultrafast, a premium speed tier for workloads where latency matters. The company reports up to 8× faster token generation in Codex and up to 6× in the API; the API guide describes the speed figure as a best-effort comparison.
Availability and plan scope
OpenAI says GPT-6 Astra Ultrafast is available in the API and in ChatGPT Work and Codex for Pro 500 and Enterprise plans. The recap describes Pro 500 as offering 25 times the ChatGPT Plus usage allowance and including Ultrafast. The public announcement does not establish that the same access or speed applies across all plans or surfaces.
Two different products, not one promise
The API speed tier and the ChatGPT plan are separate offerings. OpenAI's docs describe API limits and best-effort throughput, while the DevDay recap describes access for Work and Codex. GPT-6.1 Sol Ultrafast was listed as coming soon rather than available at announcement.