Yuan3.0 Flash: An Open Multimodal Large Language Model for Enterprise Applications
This addresses performance and efficiency issues in enterprise AI applications, offering an incremental improvement through a novel training algorithm.
The paper tackles the overthinking problem in Large Reasoning Models by introducing Yuan3.0 Flash, an open multimodal MoE model with 3.7B activated parameters, which achieves superior performance on enterprise tasks like RAG and table understanding while matching frontier models in reasoning with 1/4 to 1/2 the tokens.
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.