List your company

Visual Question Answering on GranuVistaVQA

35.74ROUGE-LQwen3-VL-8B† (GranuRAG)
9.885616.597823.3130.0222May 14, 2026

Evaluation results

MethodLinks
Qwen3-VL-8B† (GranuRAG)Setting=GranuRAG (C), Fine-tuned=true2026.0535.7446.9670.24
GranuRAGGenerator Backbone=Qwen-VL-Max2026.0532.2752.1979.3
Qwen-VL-Max (GranuRAG)Setting=GranuRAG (C)2026.0532.0151.683.9
Qwen3-VL-8B†Setting=CoT (B), Fine-tuned=true2026.0531.9646.0563.2
Qwen-VL-MaxSetting=CoT (B)2026.0530.2948.3974.1
Embedding RetrievalGenerator Backbone=Qwen-VL-Max, Retrieval Strategy=CLIP-based dense retrieval2026.0529.4745.5763.45
Qwen3-VL-8B†Setting=Baseline (A), Fine-tuned=true2026.0528.6145.3856.9
VisRAGGenerator Backbone=Qwen-VL-Max2026.0524.0643.3568.06
BaselineGenerator Backbone=Qwen-VL-Max2026.0523.7940.8356.4
Qwen-VL-MaxSetting=Baseline (A)2026.0523.3440.5354.7
GPT-4.1-Mini (GranuRAG)Setting=GranuRAG (C)2026.0522.7244.0475.1
RAVQA(PreFLMR)Generator Backbone=Qwen-VL-Max2026.0521.2742.669.24
GPT-4.1-MiniSetting=CoT (B)2026.0519.5542.7968.9
GPT-4o (GranuRAG)Setting=GranuRAG (C)2026.0519.1643.4375.4
Qwen3-VL-8B (GranuRAG)Setting=GranuRAG (C), Fine-tuned=false2026.0518.8246.4965
GPT-4.1-MiniSetting=Baseline (A)2026.0518.8140.9456.3
GPT-4oSetting=CoT (B)2026.0518.4142.0963.97
Claude-3.5-SonnetSetting=CoT (B)2026.0516.3939.5466.7
Claude-3.5-Sonnet (GranuRAG)Setting=GranuRAG (C)2026.0516.3641.2780.2
Claude-3.5-SonnetSetting=Baseline (A)2026.0515.1137.2654.5
GPT-4oSetting=Baseline (A)2026.0512.8339.9552.7
Qwen3-VL-8BSetting=CoT (B), Fine-tuned=false2026.0512.2344.4661.17
Qwen3-VL-8BSetting=Baseline (A), Fine-tuned=false2026.0510.8840.0452.9