GPT-6.1 Sol gets 8x speed at $60 per million output tokens
On October 8, OpenAI started rolling out Ultrafast mode for GPT-6.1 Sol in ChatGPT Work, Codex and the API. The mode costs $60 per million output tokens and $12 per million input, and OpenAI says it runs 8x faster than Sol Standard. That price is six times standard Sol’s $2/$10 on both input and output, so in the API every request sent on the Ultrafast tier pays the higher rate.
In the API, Ultrafast is a per-request setting: model: "gpt-6.1-sol" with service_tier: "ultrafast", and it works only in the Responses API, according to OpenAI’s model documentation for GPT-6.1 Sol. The same page says tool calling also requires the Responses API, and Chat Completions works only without tools. US and EU data residency covers the Fast and Ultrafast modes too, so teams with residency requirements can use the speed tier.
The model’s other limits apply in Ultrafast as well. reasoning.effort accepts low, medium (the default), high, xhigh and max, and the none and minimal settings are not supported. OpenAI pitches Sol as near-Astra performance at a lower cost for coding, computer use and professional work, and it tells developers to test both models on their own tasks before choosing.
At the end of September, Ultrafast was available only on GPT-6 Astra, and Sol was listed as coming soon. Back then, getting it in Codex and ChatGPT Work required the new Pro 500 plan. OpenAI has not said whether Sol’s Ultrafast comes with the same requirement.
The mode itself dates back further. TechCrunch reported that OpenAI first offered Ultrafast on August 13 as a preview on GPT-5.6 Sol. That version ran on hardware from OpenAI’s partnership with Cerebras, was open to a small group of customers, and OpenAI said it would widen access as “capacity grows.”
Up next
Claude Dashboards shows the query behind every chart, Video
“Click any number to see the query behind it,” Anthropic says of Claude Dashboards, now in beta on paid plans. What it hasn’t said, per The New Stack, is whose credentials refresh the data.
Related stories
- OpenAI puts 8x faster Codex behind a new Pro 500 plan
- Codex and ChatGPT Work give GPT-6.1 Sol 500k context
- GPT-6.1 Sol matches Astra on DeepSWE at a fifth of the cost
- ChatGPT and Codex will share one usage pool in December
- OpenAI doubles Pro and Business Premium usage limits
- Codex will spend a saved limit reset, but only when asked
Comments
No comments yet. Be the first.
Join the conversation
Sign in to leave a comment. It appears after moderation.

