Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Short-form single-page QA on NQ (Natural Questions) (official)
Loading...
50.2
EM
AGAR
9.952
20.401
30.85
41.299
Jun 11, 2026
EM
F1 Score
Average Token Count
Updated 1mo ago
Evaluation Results
Method
Method
Links
EM
F1 Score
Average Token Count
AGAR
Backbone=Qwen3-VL-8B-I...
2026.06
50.2
66.5
291
AGLA
Backbone=Qwen3-VL-8B-I...
2026.06
48.8
64.6
268
Uniform
Backbone=Qwen3-VL-8B-I...
2026.06
48.2
64.4
292
Plain VQA
Backbone=Qwen3-VL-8B-I...
2026.06
47.3
63.8
268
VEA-Br
Backbone=Qwen3-VL-8B-I...
2026.06
47.2
63.9
268
VEA-Con
Backbone=Qwen3-VL-8B-I...
2026.06
47.2
63.9
268
Random
Backbone=Qwen3-VL-8B-I...
2026.06
47.1
63.7
301
VAR
Backbone=Qwen3-VL-8B-I...
2026.06
46.9
63.6
268
CGR
Backbone=Qwen3-VL-8B-I...
2026.06
11.5
29.3
268
Feedback
Search any
task
Search any
task