OpenAI is previewing Ultrafast mode, an API tier that serves GPT-5.6 Sol at up to 14 times standard speed, launching first to a select group of customers and widening as capacity grows. It sits well above Fast mode, which tops out at 2.5x for twice the standard price.
Read the full story: OpenAI previews Ultrafast mode: GPT-5.6 Sol at 14x speed →
Transcript
OpenAI just previewed a new API tier called Ultrafast mode. It runs GPT-5.6 Sol at up to fourteen times standard speed. For context, the fast tier OpenAI shipped six weeks ago tops out at two and a half times, and costs double. So this is a big jump. But the number isn't the interesting bit. Look at how you get access. Select customers first, expanding as capacity grows. A software speedup doesn't need a waiting list. You ship it and everyone gets it at once. Capacity gating means the speed is coming from hardware, and OpenAI only has so much of it. Back in July they started serving Sol on Cerebras wafer scale chips at around fifteen times GPU speed. They haven't confirmed the link. The math lines up. Either way, latency is now something you buy.