Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Saliency-grounded Question Answering on SalBench p3 (test)
Loading...
89.2
Accuracy
GPT-4o
38.448
51.624
64.8
77.976
Apr 6, 2026
Accuracy
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy
GPT-4o
2026.04
89.2
Claude-3.5 Sonnet
2026.04
86.7
Qwen2.5-VL-7B + Saliency-R1
Training Stage=Salienc...
2026.04
63.7
Qwen2.5-VL-3B + Saliency-R1
Training Stage=Salienc...
2026.04
59.6
Qwen2.5-VL-7B + SFT
Training Stage=SFT
2026.04
57.1
Qwen2.5-VL-3B + SFT
Training Stage=SFT
2026.04
51.7
Qwen2.5-VL-7B
Training Stage=Base
2026.04
49.1
Qwen2.5-VL-3B
Training Stage=Base
2026.04
40.4
Feedback
Search any
task
Search any
task