OpenAI 为 GPT-6.1 Sol 推出 Ultrafast 模式，最高提速 8 倍

October 8, 2026, 6:45pm 1

Near-Astra intelligence at up to 8x faster speeds than Sol Standard, so you can build as fast as the ideas come.

API pricing for GPT-6.1 Sol in Ultrafast mode is $12 per million input tokens and $60 per million output tokens which is just 1.2x the cost of Astra.

Built for work where speed and intelligence make a difference: debugging an outage, agents navigating apps, and live experiences where every second counts.

In Codex and ChatGPT Work, access is available on Pro 500, eligible usage-based Enterprise, and credit-based Edu plans. Enterprise admins must enable access.

Ultrafast for GPT-6.1 Sol is available in all supported regions, including support for data residency in the US and EU. We’ve also added support for EU data residency for GPT-6.1 Sol Fast and GPT-6 Luna Fast.

Rocket emoji needed on reactions list yet again!

VeitB October 8, 2026, 7:17pm 4

Ultrafast also works with voice.
Maybe we can have rapper emoji?

Computer use is also supported.

jns2001 October 8, 2026, 7:30pm 5

Ultrafast sounds great, but most of us can’t tell whether up to 8x is worth jumping to Pro 500 without trying it first. Could you give everyone a short trial, even a few days or a capped number of runs, so we can see the speed for ourselves on real Codex work? That would make a great day in the 28.

VeitB October 8, 2026, 7:34pm 6

You can log into Codex with an API key and then call Ultrafast 6.1-Sol from there. I haven’t tried this myself yet.

config.toml

model = "gpt-6.1-sol"
service_tier = "ultrafast"

In the CLI you can use:

codex --model gpt-6.1-sol --config 'service_tier="ultrafast"'

Since I don’t know how many tokens your projects need for a few conversational turns, you might want to look into that first.

Pretty sure you need the codex 500 plan for this, I am not seeing ultrafast behavior on the 200 dollar plan (though it is not rejecting the service tier)