Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Optical Character Recognition on Lens
Loading...
68.51
Accuracy
Qwen2.5-VL
25.3708
36.5704
47.77
58.9696
May 21, 2025
Accuracy
Updated 20d ago
Evaluation Results
Method
Method
Links
Accuracy
Qwen2.5-VL
Model size=72B
2025.05
68.51
Qwen2.5-VL
Model size=32B
2025.05
65.64
Gemini2.5-Pro
Model size=-
2025.05
61.95
Qwen2.5-VL
Model size=7B
2025.05
61.65
QVQ-Max
Model size=72B
2025.05
58.67
InternVL3
Model size=78B
2025.05
54.21
Qwen2.5-VL
Model size=3B
2025.05
52.43
InternVL3
Model size=38B
2025.05
51.72
InternVL3
Model size=9B
2025.05
48.79
Deepseek-VL2
Model size=MoE 4.5B/27B
2025.05
48.76
InternVL3
Model size=2B
2025.05
45.51
Deepseek-VL2-tiny
Model size=MoE 1B/3B
2025.05
44.01
GPT-4o
Model size=-
2025.05
42.86
Gemma3
Model size=12B
2025.05
33.98
Kimi-VL-thinking
Model size=MoE 2.8B/16B
2025.05
30.21
Gemma3
Model size=4B
2025.05
27.03
Feedback
Search any
task
Search any
task