OPENAI
OpenAI turns UP the speed
OpenAI is testing a new Ultrafast mode for GPT-5.6 Sol that can run up to 14 faster than its Standard mode, starting with a limited release through the OpenAI API.
Powered by Cerebras, Ultrafast can produce up to 750 output tokens per second.
It’s designed for tasks where fast responses really matter.
Until now, getting fast AI responses often meant using a smaller model.
OpenAI says Ultrafast is meant to offer both speed and stronger AI performance.
Early uses include:
Checking logs and code while an outage is happening.
Answering questions, checking stock and fixing issues in real time.
Analysing information, testing ideas and spotting suspicious activity faster.
Done choosing between brains and speed
OpenAI is also using Ultrafast internally for engineering and research, where quicker responses can help teams test ideas and make decisions faster.
For now, Ultrafast is only available to a small group of customers, with wider access planned later.
Can’t wait for tokens to be spent in 2 minutes. - MV