Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Visual Logic on VisualPuzzles
Loading...
43.15
Top-1 Accuracy
Qwen3-VL-8B-Instruct (Teacher)
12.1684
20.2117
28.255
36.2983
May 30, 2026
Top-1 Accuracy
Top-16 Accuracy
Updated 1d ago
Evaluation Results
Method
Method
Links
Top-1 Accuracy
Top-16 Accuracy
Qwen3-VL-8B-Instruct (Teacher)
Model Size=8B
2026.05
43.15
-
Standard On-Policy Distillation
Student Model=Qwen3-VL...
2026.05
40.75
39.17
VGS On-Policy Distillation
Student Model=Qwen3-VL...
2026.05
40.31
39.22
VGS On-Policy Distillation
Student Model=Qwen3-VL...
2026.05
31.76
31.59
Standard On-Policy Distillation
Student Model=Qwen3-VL...
2026.05
28.08
30.64
Qwen3-VL-4B-Instruct (Initial Student)
Model Size=4B
2026.05
26.54
-
Qwen3-VL-2B-Instruct (Initial Student)
Model Size=2B
2026.05
13.36
-
Feedback
Search any
task
Search any
task