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

Visual Question Answering on 14 benchmarks 8 spatial and 6 real-world (test)

74.5Average Accuracy (14 Benchmarks)

SpatialThinker-30B

56.09260.87165.6570.429Nov 10, 2025
Updated 18d ago

Evaluation Results

MethodLinks
2025.11
74.56.438.75.8
2025.11
71.5----
2025.11
70.57.7-14.71.8
2025.11
68.7----
2025.11
68.1----
2025.11
67.24.4-4.31.4-1.5
2025.11
65.99.1-5.60.1-2.8
2025.11
65.8----
2025.11
64.92.1-6.6-0.9-3.8
2025.11
62.8----
2025.11
625.2-9.5-3.8-6.7
2025.11
60.9----
2025.11
60.73.9-10.8-5.1-8
2025.11
56.8----