Megrez-Omni Technical Report

Boxun Li Yadong Li Zhiyuan Li Congyi Liu Weilin Liu Guowei Niu, Zheyue Tan Haiyang Xu Zhuyu Yao Tao Yuan Dong Zhou Yueqing Zhuang, Shengen Yan Guohao Dai Yu Wang · 2025

Read Megrez-Omni Technical Report by Boxun Li Yadong Li Zhiyuan Li Congyi Liu Weilin Liu Guowei Niu, Zheyue Tan Haiyang Xu Zhuyu Yao Tao Yuan Dong Zhou Yueqing Zhuang, Shengen Yan Guohao Dai Yu Wang (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-15803.txt.

Description

Boxun Li and co-authors present a pair of compact AI models built for fast use on edge devices, in a 2025 technical report. One model handles text alone, and the other adds image and audio understanding. Both hold three billion parameters, a size that fits phones and small machines.

The text model is tuned for accuracy, speed, and ease of use across many tasks. The omni version builds on it to read images, text, and sound in one system. The authors claim top scores among peers across all three modes. Software and hardware were designed together to keep response times low. Such models suit apps that cannot rely on the cloud.

Catalog record

CREATORBoxun Li Yadong Li Zhiyuan Li Congyi Liu Weilin Liu Guowei Niu, Zheyue Tan Haiyang Xu Zhuyu Yao Tao Yuan Dong Zhou Yueqing Zhuang, Shengen Yan Guohao Dai Yu Wang
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE36 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2502-15803

Browse the full library · About the Tower of Babel