List your company
ResearchFormerly WizWand

Financial Multimodal Reasoning on FinMME

49.64AccuracyQwen2.5-VL
15.080824.052933.02541.9971Dec 9, 2025Jan 2, 2026Jan 26, 2026Feb 19, 2026Mar 15, 2026Apr 8, 2026May 2, 2026
Updated 5mo ago

Evaluation results

MethodLinks
Qwen2.5-VLParameter Scale=72B2025.1249.64
Qwen2-VLParameter Scale=72B2025.1237.11
SYNDATAModel=Qwen3-VL-4B, Data Size=40002026.0536.96
DeepSeekVL-2 Small2025.1234.14
Qwen2-VLParameter Scale=7B2025.1234.14
SYNDATAModel=Qwen3-VL-4B, Data Size=10002026.0533.81
SYNDATAModel=Qwen3-VL-4B, Data Size=20002026.0533.46
Qwen2.5 VL-7B + PHMCompression Method=PHM, Parameter Scale=5.38B2025.1233.1
Qwen2.5-VLParameter Scale=3B2025.1232.53
SYNDATAModel=Qwen2.5-VL-7B, Data Size=40002026.0532.47
LLaVA-1.5-7B + PHMCompression Method=PHM, Parameter Scale=5.15B2025.1228.5
Base ModelModel=Qwen3-VL-4B, Data Size=-2026.0527.65
SYNDATAModel=Qwen2.5-VL-7B, Data Size=20002026.0526.98
InstructBLIPBackbone=Vicuna-13B2025.1226.21
SYNDATAModel=Qwen2.5-VL-7B, Data Size=10002026.0524.9
InstructBLIPBackbone=Vicuna-7B2025.1224.5
InstructBLIP + PHMCompression Method=PHM, Parameter Scale=5.45B2025.1221.5
Base ModelModel=Qwen2.5-VL-7B, Data Size=-2026.0516.41