Get the latest tech news

MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second


MiMo, in collaboration with TileRT, releases the UltraSpeed mode of Xiaomi MiMo-V2.5-Pro — breaking 1000 tokens/s generation speed on a 1T-parameter model for the first time on commodity GPUs through extreme model-system codesign.

None

Get the Android app

Or read this on Hacker News

Read more on:

Photo of tokens

tokens

Photo of mimo

mimo

Photo of T model

T model

Related news:

News photo

What Are Tokens in LLMs?

News photo

Tokenomics: Quantifying Where Tokens Are Used in Agentic Software Engineering

News photo

Show HN: Lowfat – pluggable CLI filter that saved 91.8% of my LLM tokens