Multimodal Document Question Answering on LongDocURL
65.7Overall AccHIEVI-RAG
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| HIEVI-RAGParam.=8B2026.07 | 65.7 | — | — | — | — | — | — | — | |
| OursCategory=RAG-based, Top-k retrieval=Top-42026.01 | 60.7 | — | — | — | — | — | — | — | |
| OursCategory=RAG-based, Top-k retrieval=Top-12026.01 | 56.8 | — | — | — | — | — | — | — | |
| Doc-V*Param.=7B, Optimization=GRPO-optimized2026.07 | 56.3 | — | — | — | — | — | — | — | |
| MDocAgentCategory=RAG-based, Top-k retrieval=Top-42026.01 | 56.1 | — | — | — | — | — | — | — | |
| ALDENParam.=7B, Pipeline=ColQwen+ColBERT2026.07 | 55.1 | — | — | — | — | — | — | — | |
| URaGParam.=7B2026.07 | 52.2 | — | — | — | — | — | — | — | |
| MoLoRAG+Param.=7B2026.07 | 51.9 | — | — | — | — | — | — | — | |
| MLDocRAGRetriever=BGE-m3, Generator=LVLM, Modality (TXT)=true, Modality (IMG)=true2026.02 | 50.8 | 65.1 | 39.4 | 48.3 | 41.1 | 66.9 | 56.3 | 23.4 | |
| M3DocRAGParam.=7B2026.07 | 49 | — | — | — | — | — | — | — | |
| RAG^txt+txt2i_BM25Retriever=BM25, Generator=LLM, Modality (TXT)=true, Modality (IMG)=false2026.02 | 48.1 | 60.6 | 36.5 | 45.7 | 40.3 | 61.9 | 54.1 | 22.4 | |
| RAG^txt+txt2i_BGE-m3Retriever=BGE-m3, Generator=LLM, Modality (TXT)=true, Modality (IMG)=false2026.02 | 47.8 | 59.7 | 32.8 | 47 | 42.3 | 61.9 | 57.4 | 17.1 | |
| VRAG^page_ColPaliRetriever=ColPali, Generator=LVLM, Modality (TXT)=false, Modality (IMG)=true2026.02 | 47.1 | 56.8 | 39.4 | 45.8 | 53.1 | 52.1 | 50.3 | 36.4 | |
| MDocAgentParam.=7B2026.07 | 46.9 | — | — | — | — | — | — | — | |
| GraphRAG^txt+txt2i_BGE-m3Retriever=BGE-m3, Generator=LLM, Modality (TXT)=true, Modality (IMG)=false2026.02 | 45.3 | 58.2 | 33.9 | 32.8 | 36.1 | 59.2 | 54.4 | 15.7 | |
| LC^txt+imgRetriever=-, Generator=LVLM, Modality (TXT)=true, Modality (IMG)=true2026.02 | 40.7 | 57.7 | 34.7 | 30.9 | 39.8 | 46 | 48.9 | 23.4 | |
| LC^txtRetriever=-, Generator=LLM, Modality (TXT)=true, Modality (IMG)=false2026.02 | 40.2 | 59.1 | 38.8 | 26 | 41.5 | 43.3 | 45.7 | 29.2 | |
| Qwen-2.5-VL-7B-InstructCategory=LVLM-based2026.01 | 39.8 | — | — | — | — | — | — | — | |
| RAG^txt_BM25Retriever=BM25, Generator=LLM, Modality (TXT)=true, Modality (IMG)=false2026.02 | 39.5 | 60.2 | 37.4 | 24.9 | 37.3 | 43.1 | 45.9 | 26.8 | |
| MRAG^txt+img_ColPaliRetriever=ColPali, Generator=LVLM, Modality (TXT)=true, Modality (IMG)=true2026.02 | 38.9 | 51 | 31.8 | 32.9 | 34 | 45.2 | 44.8 | 23.4 | |
| LC^txt+txt2iRetriever=-, Generator=LLM, Modality (TXT)=true, Modality (IMG)=false2026.02 | 37.3 | 51.1 | 31 | 29.1 | 33.6 | 42.2 | 49 | 15.7 | |
| RAG^txt_BGE-m3Retriever=BGE-m3, Generator=LLM, Modality (TXT)=true, Modality (IMG)=false2026.02 | 36.8 | 59.8 | 34.7 | 19.3 | 31.5 | 41.4 | 44.6 | 21 | |
| MRAG^txt+img_CLIPRetriever=CLIP, Generator=LVLM, Modality (TXT)=true, Modality (IMG)=true2026.02 | 34.5 | 51.7 | 32.4 | 22.6 | 30.7 | 35.6 | 44 | 21 | |
| MRAG^txt+img_SigLIPRetriever=SigLIP, Generator=LVLM, Modality (TXT)=true, Modality (IMG)=true2026.02 | 34.5 | 51.7 | 32.4 | 22.6 | 30.7 | 35.6 | 44 | 21 | |
| LC^pageRetriever=-, Generator=LVLM, Modality (TXT)=false, Modality (IMG)=true2026.02 | 31.3 | 40.2 | 29.7 | 23.7 | 30.3 | 34 | 39.1 | 17.9 | |
| VRAG^page_CLIPRetriever=CLIP, Generator=LVLM, Modality (TXT)=false, Modality (IMG)=true2026.02 | 26.2 | 33.8 | 26.2 | 20.1 | 26.1 | 24.1 | 34.2 | 18.9 | |
| VRAG^page_SigLIPRetriever=SigLIP, Generator=LVLM, Modality (TXT)=false, Modality (IMG)=true2026.02 | 19.9 | 21.2 | 25.3 | 17 | 19.1 | 17.4 | 17.7 | 25.8 | |
| LLaVA-1.6-7B-InstructCategory=LVLM-based2026.01 | 11 | — | — | — | — | — | — | — |