# Multimodal Document Question Answering on DocBench (test) Source: [Multimodal Document Question Answering on DocBench (test)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:34:05.842Z Evaluation of Multimodal Document Question Answering performance on the test split of the DocBench dataset, reporting domain-specific accuracy scores across categories like Academic, Finance, Government, Law, and News. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Multimodal Document Question Answering | | Dataset (including any stated split) | DocBench (test) | | Metric | Accuracy (Academic) | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 6 of 6 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Accuracy (Academic)](): Higher is better; unit: PERCENT. - [Accuracy (Finance)](): Higher is better; unit: PERCENT. - [Accuracy (Government)](): Higher is better; unit: PERCENT. - [Accuracy (Law)](): Higher is better; unit: PERCENT. - [Accuracy (News)](): Higher is better; unit: PERCENT. - [Accuracy (Overall)](): Higher is better; unit: PERCENT. ## Recorded results | Method | Accuracy (Academic) | Result date | Report paper | | --- | --- | --- | --- | | BayesRAG | 52.4 | Not recorded | 6965d19b6343fa78c83236ae | | RAGAnything | 51.4 | Not recorded | 6965d19b6343fa78c83236ae | | RAGFlow | 45.2 | Not recorded | 6965d19b6343fa78c83236ae | | GPT-4o-mini | 42.5 | Not recorded | 6965d19b6343fa78c83236ae | | ViDoRAG | 29 | Not recorded | 6965d19b6343fa78c83236ae | | VisRAG | 21.9 | Not recorded | 6965d19b6343fa78c83236ae | ### BayesRAG Accuracy (Academic): 52.4. Result ID: 69968e275698665ec4d6fa3a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Response Generator":"GPT-4o-mini","Parsing":"MinerU"}. This record does not assert independent reproduction. ### RAGAnything Accuracy (Academic): 51.4. Result ID: 69968e275698665ec4d6fa39. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Response Generator":"GPT-4o-mini","Paradigm":"Knowledge-graph-based"}. This record does not assert independent reproduction. ### RAGFlow Accuracy (Academic): 45.2. Result ID: 69968e275698665ec4d6fa38. [Reporting paper]() Recorded evaluation configuration: {"Response Generator":"GPT-4o-mini","Paradigm":"Modular RAG"}. This record does not assert independent reproduction. ### GPT-4o-mini Accuracy (Academic): 42.5. Result ID: 69968e275698665ec4d6fa35. [Reporting paper]() Recorded evaluation configuration: {"Input":"Page-level screenshots (max 50)","Direct evaluation":"true"}. This record does not assert independent reproduction. ### ViDoRAG Accuracy (Academic): 29. Result ID: 69968e275698665ec4d6fa37. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Response Generator":"GPT-4o-mini","Agentic Workflow":"Explore, Summarize, Reflect"}. This record does not assert independent reproduction. ### VisRAG Accuracy (Academic): 21.9. Result ID: 69968e275698665ec4d6fa36. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Response Generator":"GPT-4o-mini","Parsing":"Visual-centric (no OCR)"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.