List your company
ResearchFormerly WizWand

Optical Character Recognition on Lens

68.51AccuracyQwen2.5-VL
25.370836.570447.7758.9696May 21, 2025
Updated 6mo ago

Evaluation results

MethodLinks
Qwen2.5-VLModel size=72B2025.0568.51
Qwen2.5-VLModel size=32B2025.0565.64
Gemini2.5-ProModel size=-2025.0561.95
Qwen2.5-VLModel size=7B2025.0561.65
QVQ-MaxModel size=72B2025.0558.67
InternVL3Model size=78B2025.0554.21
Qwen2.5-VLModel size=3B2025.0552.43
InternVL3Model size=38B2025.0551.72
InternVL3Model size=9B2025.0548.79
Deepseek-VL2Model size=MoE 4.5B/27B2025.0548.76
InternVL3Model size=2B2025.0545.51
Deepseek-VL2-tinyModel size=MoE 1B/3B2025.0544.01
GPT-4oModel size=-2025.0542.86
Gemma3Model size=12B2025.0533.98
Kimi-VL-thinkingModel size=MoE 2.8B/16B2025.0530.21
Gemma3Model size=4B2025.0527.03