Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Real-world Multimodal Interaction on RealWorldQA (test)

77.8Accuracy

Qwen2-VL-72B

59.39264.17168.9573.729Dec 6, 2024
Updated 1mo ago

Evaluation Results

MethodLinks
2024.12
77.8
2024.12
76.5
2024.12
73.8
2024.12
72.7
2024.12
71.9
2024.12
71.3
2024.12
70.7
70.4
2024.12
70.1
2024.12
69.9
2024.12
67.8
2024.12
66.3
2024.12
65.5
2024.12
65
2024.12
64.4
2024.12
64.2
2024.12
63.3
60.1