List your company
ResearchFormerly WizWand

Document Understanding on DocVQA (test)

96.5AccuracyQwen2-VL-72B
21.140.67560.2579.825Aug 6, 2024Oct 26, 2024Jan 16, 2025Apr 8, 2025Jun 28, 2025Sep 18, 2025Dec 9, 2025
Updated 8mo ago

Evaluation results

MethodLinks
Qwen2-VL-72BModel Scale=72B2024.1296.5
Claude-3.5-Sonnet2024.1295.2
Qwen2-VL-7B-InsModel Scale=7B2024.1294.5
InternVL2-76BModel Scale=76B2024.1294.1
Qwen2.5VL-3BComplexity=O(n)2025.1293.9
MAmmoTH-VL-8BModel Scale=8B, Single-image variant (SI)=true2024.1293.8
MAmmoTH-VL-8BModel Scale=8B2024.1293.7
Gemini-1.5 Pro2024.1093.1
Gemini-1.5-Pro2024.1293.1
GPT-4o2024.0892.8
GPT-4o2024.1092.8
GPT-4o2024.1292.8
ARIA2024.1092.6
Molmo-7B-DModel Scale=7B2024.1292.2
LLaVA-OV-72BModel Scale=72B, Single-image variant (SI)=true2024.1291.8
InfiniteVL-4BComplexity=O(1)2025.1291.7
InternVL-2-8BModel Scale=8B2024.1291.6
InternVL2.5-4BComplexity=O(n)2025.1291.6
LLaVA-OneVision-72BModel Scale=72B2024.0891.3
LLaVA-OV-72BModel Scale=72B2024.1291.3
INXComp-2.5-7BModel Scale=2.5B2024.1290.9
MiniCPM-V-2.6-7BModel Scale=7B2024.1290.8
Pixtral-12B2024.1090.7
Gemini-1.5 Flash2024.1089.9
GPT-4VVersion=V-Preview2024.0888.4
Llama3.2-11B2024.1088.4
GPT-4V2024.1088.4
Llama-3.2-11B-Vision-InsModel Scale=11B2024.1288.4
LLaVA-OneVision-7BModel Scale=7B2024.0887.5
LLaVA-OV-7BModel Scale=7B2024.1287.5
LLaVA-OV-7BModel Scale=7B, Single-image variant (SI)=true2024.1286.9
SmolVLM2-2BComplexity=O(n)2025.1280
Cambrian-1-8BModel Scale=8B2024.1277.8
PaliGemma2-3BComplexity=O(n)2025.1271.6
LLaVA-OneVision-0.5BModel Scale=0.5B2024.0870
Phi-3.5-Vision-4BComplexity=O(n)2025.1269.3
TinyLLaVA-3BComplexity=O(n)2025.1234.7
MaTVLM-3BComplexity=O(n)2025.1233
Cobra-3BComplexity=O(1)2025.1224