Document Question Answering on MMLongBench-Doc
65.8AccuracyHuman Baseline
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Human Baseline2026.01 | 65.8 | 66 | — | — | — | — | |
| Evolved AgentModel=Gemini-3.1-Flash, Evolved=true2026.06 | 62 | — | — | — | — | — | |
| MARDocBackbone=Qwen3-VL-32B2026.06 | 59.9 | 57.5 | — | — | — | — | |
| Evolved Agent, No Ground TruthModel=Gemini-3.1-Flash, Evolved=true2026.06 | 58.4 | — | — | — | — | — | |
| Evolved Agent, No ExplorationModel=Gemini-3.1-Flash, Evolved=true2026.06 | 57.4 | — | — | — | — | — | |
| DocAgentModel=Claude-3.5-Sonnet2026.01 | 57.3 | 54.1 | — | — | — | — | |
| MARDocBackbone=Qwen3-VL-30B-A3B2026.06 | 57.1 | 54.6 | — | — | — | — | |
| DocDancerModel=GPT-5.22026.01 | 57 | 56.8 | 67.6 | — | — | — | |
| SimpleDocModel=Gemini-3.1-Flash, Evolved=false2026.06 | 56.6 | — | — | — | — | — | |
| DocDancerModel=Gemini-2.5-Pro2026.01 | 56.3 | 55.3 | 65.9 | — | — | — | |
| Evolved Agent, GPT-5-minimal Meta AgentModel=Gemini-3.1-Flash, Evolved=true2026.06 | 56.1 | — | — | — | — | — | |
| Evolved Agent, No ExplorationModel=GPT-5-mini, Evolved=true2026.06 | 55.5 | — | — | — | — | — | |
| Evolved AgentModel=GPT-5-mini, Evolved=true2026.06 | 55.4 | — | — | — | — | — | |
| Evolved Agent, No Ground TruthModel=GPT-5-mini, Evolved=true2026.06 | 54.9 | — | — | — | — | — | |
| DocDancerModel=Qwen3-30B-A3B (ft)2026.01 | 54.4 | 53.9 | 65.3 | — | — | — | |
| MACTModel=Gemini-3.1-Flash, Evolved=false2026.06 | 54.4 | — | — | — | — | — | |
| Evolved Agent, GPT-5-minimal Meta AgentModel=GPT-5-mini, Evolved=true2026.06 | 52.8 | — | — | — | — | — | |
| DocDancerModel=GPT-4o2026.01 | 52.3 | 50.8 | 59.2 | — | — | — | |
| DocAgentModel=GPT-4o2026.01 | 51.8 | 49.1 | — | — | — | — | |
| SimpleDocModel=GPT-5-mini, Evolved=false2026.06 | 51.3 | — | — | — | — | — | |
| MDocAgentModel=Gemini-3.1-Flash, Evolved=false2026.06 | 51.2 | — | — | — | — | — | |
| DocDancerModel=Qwen3-4B (ft)2026.01 | 48.4 | 49.2 | 59.4 | — | — | — | |
| MACTModel=MiMo-VL-7B2026.01 | 47.4 | — | — | — | — | — | |
| MACTModel=GPT-5-mini, Evolved=false2026.06 | 47.4 | — | — | — | — | — | |
| File APIModel=Gemini-3.1-Flash, Evolved=false2026.06 | 45.4 | — | — | — | — | — | |
| File APIModel=GPT-5-mini, Evolved=false2026.06 | 44.6 | — | — | — | — | — | |
| MDocAgentModel=GPT-5-mini, Evolved=false2026.06 | 44 | — | — | — | — | — | |
| Naive VLModel=GPT-4o2026.01 | 42.8 | 44.9 | — | — | — | — | |
| RAGAnythingModel=GPT-4o-mini2026.01 | 42.8 | — | — | — | — | — | |
| Baseline AgentModel=Gemini-3.1-Flash, Evolved=false2026.06 | 42.4 | — | — | — | — | — | |
| MDocAgentModel=GPT-4o2026.01 | 42 | — | — | — | — | — | |
| Baseline AgentModel=GPT-5-mini, Evolved=false2026.06 | 40.7 | — | — | — | — | — | |
| CoTModel=Gemini-3.1-Flash, Evolved=false2026.06 | 40.4 | — | — | — | — | — | |
| TesseractModel=Gemini-2.0-Flash2026.01 | 39.6 | 37.2 | — | — | — | — | |
| Doc-ReactModel=GPT-4o2026.01 | 38.1 | 38.3 | — | — | — | — | |
| CoTModel=GPT-5-mini, Evolved=false2026.06 | 36.2 | — | — | — | — | — | |
| ColpaliModel=GPT-4o2026.01 | 32.2 | 30.8 | — | — | — | — | |
| M3DocRAG w/ ColPaliModel=Qwen2-VL-7B2026.01 | 31.4 | 36.5 | — | — | — | — | |
| TesseractModel=GPT-4o2026.01 | 30.1 | 30.5 | — | — | — | — | |
| VisRAGModel=GPT-4o2026.01 | 29 | 27.8 | — | — | — | — | |
| DocLensModel=Claude-4-Sonnet2026.01 | — | — | 63.3 | — | — | — | |
| DocLensModel=Gemini-2.5-Pro2026.01 | — | — | 67.6 | — | — | — | |
| GPT-4o-200b-128Retrieval=Ground-Truth2025.10 | — | — | — | 69.6 | 69.3 | 56.5 | |
| GPT-4o-200b-128Retrieval=retrieving@52025.10 | — | — | — | 57.5 | 60.7 | 30.3 | |
| GPT-4o-200b-128Retrieval=retrieving@102025.10 | — | — | — | 61 | 63.7 | 37.2 | |
| GPT-4o-200b-128Retrieval=topk-adjusted2025.10 | — | — | — | 59.3 | 62.9 | 40.9 | |
| GPT-4o-200b-128Retrieval=LAD-RAG2025.10 | — | — | — | 62.5 | 67.6 | 45 | |
| Idefics2-8BRetrieval=All-Pages2025.10 | — | — | — | 7 | 7.7 | 7.2 | |
| InternVL2-8BRetrieval=Ground-Truth2025.10 | — | — | — | 39.9 | 50.6 | 25 | |
| InternVL2-8BRetrieval=retrieving@52025.10 | — | — | — | 28.7 | 37.2 | 16.4 | |
| InternVL2-8BRetrieval=retrieving@102025.10 | — | — | — | 31.9 | 39.5 | 20.8 | |
| InternVL2-8BRetrieval=topk-adjusted2025.10 | — | — | — | 30.4 | 36.5 | 21.2 | |
| InternVL2-8BRetrieval=LAD-RAG2025.10 | — | — | — | 44.8 | 49.5 | 24.2 | |
| MiniCPM-Llama3-V2.5-8BRetrieval=All-Pages2025.10 | — | — | — | 8.5 | 9.5 | 9.5 | |
| mPLUG-DocOwl 1.5-8BRetrieval=All-Pages2025.10 | — | — | — | 6.9 | 7.4 | 6.4 | |
| Naive VLModel=Gemini-2.5-Pro2026.01 | — | — | 58.1 | — | — | — | |
| Phi-3.5-Vision-4BRetrieval=Ground-Truth2025.10 | — | — | — | 38.3 | 49.2 | 22.7 | |
| Phi-3.5-Vision-4BRetrieval=retrieving@52025.10 | — | — | — | 31.5 | 39.9 | 19.2 | |
| Phi-3.5-Vision-4BRetrieval=retrieving@102025.10 | — | — | — | 30.2 | 38.1 | 18.9 | |
| Phi-3.5-Vision-4BRetrieval=topk-adjusted2025.10 | — | — | — | 33.6 | 40 | 19.9 | |
| Phi-3.5-Vision-4BRetrieval=LAD-RAG2025.10 | — | — | — | 39.1 | 41.4 | 20.2 | |
| Pixtral-12B-2409Retrieval=Ground-Truth2025.10 | — | — | — | 49.8 | 53.7 | 34.3 | |
| Pixtral-12B-2409Retrieval=retrieving@52025.10 | — | — | — | 38.9 | 39.5 | 20 | |
| Pixtral-12B-2409Retrieval=retrieving@102025.10 | — | — | — | 34.5 | 32.7 | 17.8 | |
| Pixtral-12B-2409Retrieval=topk-adjusted2025.10 | — | — | — | 38.3 | 39.4 | 21.3 | |
| Pixtral-12B-2409Retrieval=LAD-RAG2025.10 | — | — | — | 41.5 | 46.2 | 26.1 | |
| SimpleDocModel=Claude-4-Sonnet2026.01 | — | — | 58.6 | — | — | — | |
| SimpleDocModel=Gemini-2.5-Pro2026.01 | — | — | 56.6 | — | — | — |