Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Needle-in-a-haystack localization on Pinpoint-Bench
Loading...
55.2
Accuracy
PixelEyes
28.7736
35.6343
42.495
49.3557
Jun 30, 2026
Accuracy
TAE
LSR
Updated 23d ago
Evaluation Results
Method
Method
Links
Accuracy
TAE
LSR
PixelEyes
Size=8B
2026.06
55.2
26.64
74.83
PixelEyes
Size=4B
2026.06
54.73
26.13
76.91
Qwen-3-VL
Size=8B
2026.06
49.88
-
-
Qwen-3-VL
Size=4B
2026.06
46.19
-
-
Mini-o3
Size=7B
2026.06
44.34
8.38
78.52
Gemini-3-Flash
Size=-
2026.06
42.26
-
-
Thyme
Size=7B
2026.06
40.42
-
-
DeepEyes
Size=7B
2026.06
39.72
14.89
20.79
Qwen-2.5-VL
Size=7B
2026.06
39.03
-
-
Pixel-Reasoner
Size=7B
2026.06
29.79
15.56
46.88
Feedback
Search any
task
Search any
task