Yuan3.0 Flash: An Open Multimodal Large Language Model for Enterprise Applications

TL;DR

Yuan3.0 Flash is an open-source multimodal large language model optimized for enterprise tasks, featuring a novel RL training method to reduce overthinking and achieving high performance with fewer tokens.

Contribution

It introduces Yuan3.0 Flash, a multimodal LLM with a new RL training algorithm (RAPO) to improve enterprise task performance and reduce overthinking, while being openly available for research.

Findings

Superior performance on enterprise tasks like RAG and summarization.

Achieves high reasoning accuracy with fewer tokens.

Open-sourced for community research and deployment.

Abstract

We introduce Yuan3.0 Flash, an open-source Mixture-of-Experts (MoE) MultiModal Large Language Model featuring 3.7B activated parameters and 40B total parameters, specifically designed to enhance performance on enterprise-oriented tasks while maintaining competitive capabilities on general-purpose tasks. To address the overthinking phenomenon commonly observed in Large Reasoning Models (LRMs), we propose Reflection-aware Adaptive Policy Optimization (RAPO), a novel RL training algorithm that effectively regulates overthinking behaviors. In enterprise-oriented tasks such as retrieval-augmented generation (RAG), complex table understanding, and summarization, Yuan3.0 Flash consistently achieves superior performance. Moreover, it also demonstrates strong reasoning capabilities in domains such as mathematics, science, etc., attaining accuracy comparable to frontier model while requiring only approximately 1/4 to 1/2 of the average tokens. Yuan3.0 Flash has been fully open-sourced to facilitate further research and real-world deployment: https://github.com/Yuan-lab-LLM/Yuan3.0.