README.md

metadata

license: other
library_name: transformers
pipeline_tag: image-text-to-text

Yuan 3.0 Multimodal Foundation Model

---

---

This repository contains Yuan 3.0 Flash, a 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. It was introduced in the paper Yuan3.0 Flash: An Open Multimodal Large Language Model for Enterprise Applications.

Latest Updates 🎉🎉

• [2025-12-30] Released Yuan 3.0-40B Multimodal Large Language Model, a high-performance model for enterprise-grade application scenarios: Yuan3.0 Flash

1. Introduction

Yuan 3.0 Flash, developed by the YuanLab.ai team, is a 40B parameter multimodal foundation model that employs a Mixture of Experts (MoE) architecture, activating only approximately 3.7B parameters per inference. Through innovative reinforcement learning training methods (RAPO), it significantly reduces inference token consumption while improving reasoning accuracy, exploring the innovative path of "less computation, higher intelligence" for large language models.

Fig.1: Yuan3.0 Multimodal Large Language Model Architecture

Core Features

• 🚀 Efficient Inference: Reduces inference token consumption by up to 75%, significantly lowering costs
• 🎯 Enterprise-Grade Optimization: Deeply optimized for enterprise scenarios such as RAG, document understanding, and table analysis
• 🎨 Multimodal Support: Supports text, image, table, document and other multimodal inputs
• 📚 Long Context: Supports 128K context length, achieving 100% accuracy in "Needle in a Haystack" tests
• ⚡ Ready-to-Use Intelligence: Default inference mode meets the needs of most enterprise scenarios

2. Performance

Yuan 3.0 Flash outperforms GPT-5.1 in enterprise-grade RAG, multimodal retrieval, table understanding, summary generation and other tasks. With 40B parameters, it achieves the reasoning accuracy of 235B/671B models while reducing token consumption by 50%-75%, providing enterprises with high-performance, low-cost large language model solutions.

Fig.2: Yuan3.0 Flash Evaluation Results

3. Core Technology

RAPO Reinforcement Learning Algorithm

The innovative Reflection-aware Adaptive Policy Optimization (RAPO) algorithm, through the Reflection Inhibition Reward Mechanism (RIRM):

• ✅ Identifies the key point where the correct answer is first obtained
• 🎯 Suppresses subsequent redundant reasoning behavior
• 📉 Improves accuracy while reducing inference token count by approximately 75%

Training Method	AIME 2024 Accuracy	Avg Output Length	MATH-500 Accuracy	Avg Output Length
Yuan3.0 Flash (40B) SFT	31.45%	13,656 tokens	83.20%	3,362 tokens
RL+DAPO length-penalty	46.35%	13,781 tokens	89.06%	3,974 tokens
RL+RIRM	47.92%	7,505 tokens	89.47%	1,777 tokens

4. Model Download

We provide download links for multiple model formats:

Model	Parameters	Precision	Sequence Length	Model Format	Download Link
Yuan3.0 Flash	40B	16bit	128K	HuggingFace	ModelScope | HuggingFace | WiseModel
Yuan3.0 Flash 4bit	40B	4bit	128K	HuggingFace	ModelScope | HuggingFace | WiseModel

5. Evaluation Results

5.1 Text-based RAG Evaluation: ChatRAG 🏆

Yuan 3.0 Flash leads DeepSeek-V3, DeepSeek-R1 and other large language models in average accuracy across 10 evaluation tasks in the industry-standard RAG benchmark ChatRAG.

Model Average Accuracy Comparison

Models	Avg All	D2D	QuAC	QReCC	CoQA	DoQA	CFQA	SQA	TCQA	HDial	INSCIT
DeepSeek-V3	50.47	31.59	28.86	49.31	76.98	26.11	83.49	82.13	46.69	47.43	32.08
OpenAI GPT-4o	50.54	32.76	26.56	49.30	76.11	28.78	81.85	81.14	49.75	41.29	26.69
Yuan3.0 Flash	64.47	49.82	53.79	57.08	90.93	59.99	74.40	87.52	66.31	68.45	36.40

---

5.2 Multimodal RAG Evaluation: Docmatix 🏆

Models	Avg.
Qwen2.5-VL-72B-Instruct	59.75
OpenAI GPT-4V	60.10
Yuan3.0 Flash	65.07

---

5.3 Multimodal Complex Table Content Analysis Evaluation: MMTab 🏆

Models	Avg.	TABMWP	WTQ	WTQ	HiTab
OpenAI GPT-5.1	55.15	64.95	60.77	77.77	61.37
Yuan3.0 Flash	58.29	95.09	68.23	69.80	69.17

---

5.4 Text Summarization Generation Evaluation: SummEval 🏆

Models	Avg.	Lexical Overlap ROUGE-1	Semantic Similarity BERTScore	Factual Consistency SummaC
DeepSeek-V3	59.28	25.50	86.30	68.20
Yuan3.0 Flash	59.31	51.32	89.99	45.34

6. Quick Start

For specific usage methods, please refer to the official QuickStart guide.

7. License Agreement

The use of Yuan 3.0 code and models must comply with the 《Yuan 3.0 Model License Agreement》. The Yuan 3.0 model supports commercial use without requiring authorization application.

8. Citation

@article{yuan3flash2025,
  title={Yuan3.0 Flash: An Open Multimodal Large Language Model for Enterprise Applications},
  author={YuanLab.ai and others},
  journal={arXiv preprint arXiv:2601.01718},
  year={2025}
}