Celeris launches a lab to build an LLM with microsecond response times

timshi_ai · x · 2026-07-23

Celeris is being launched as a new AI research lab focused on building the world’s fastest LLM.

The team says it is working on algorithmic and engineering advances to deliver frontier intelligence at microsecond response times, aiming at workloads like:

Its pitch is that today’s builders often have to choose smaller models or ration frontier intelligence because the latency is too high.

Original post →

More from Companies & People

Companies & People channel →