Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Science Question Answering on SQA
Loading...
98.76
Exact Match
Baseline
73.8416
80.3108
86.78
93.2492
Feb 10, 2026
Feb 17, 2026
Feb 24, 2026
Mar 3, 2026
Mar 10, 2026
Mar 17, 2026
Mar 24, 2026
Exact Match
Updated 24d ago
Evaluation Results
Method
Method
Links
Exact Match
Baseline
2026.02
98.76
IDPruner
Retain Tokens=25%
2026.02
95.14
VisionSelector
Retain Tokens=25%
2026.02
94.7
DART
Retain Tokens=25%
2026.02
93.11
IDPruner
Retain Tokens=10%
2026.02
88.7
VISOR
Backbone=LLaVA-OV 1.5B...
2026.03
84.7
Downsample
Backbone=LLaVA-OV 1.5B...
2026.03
76.9
VisionZip†
Backbone=LLaVA-OV 1.5B...
2026.03
76.2
PyramidDrop
Backbone=LLaVA-OV 1.5B...
2026.03
76
VisPruner
Backbone=LLaVA-OV 1.5B...
2026.03
76
HiRED
Backbone=LLaVA-OV 1.5B...
2026.03
76
LLaVA-OV
Backbone=LLaVA-OV 1.5B...
2026.03
75.6
VisionZip
Backbone=LLaVA-OV 1.5B...
2026.03
74.8
SparseVLM
Backbone=LLaVA-OV 1.5B...
2026.03
74.8
Feedback
Search any
task
Search any
task