Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Multimodal Preference Evaluation on Polaris
Loading...
57.8
tau_c
Polos
52.08
53.565
55.05
56.535
Oct 1, 2025
tau_c
P-Acc
Updated 17d ago
Evaluation Results
Method
Method
Links
tau_c
P-Acc
Polos
reference-based=true
2025.10
57.8
-
RefPAC-S
reference-based=true
2025.10
56
-
LongCLIP-S
reference-based=false
2025.10
54
77.5
BLIP-S
reference-based=false
2025.10
54
79.5
VL-GUIDE-S-VLM
reference-based=false
2025.10
53.9
79.4
PAC-S
reference-based=false
2025.10
52.5
77
CLIP-S
reference-based=false
2025.10
52.3
79.7
Ref-free Polos
reference-based=false
2025.10
52.3
60
RefCLIP-S
reference-based=true
2025.10
52.3
-
ImgREW-S
reference-based=false
2025.10
52.3
73.3
Feedback
Search any
task
Search any
task