Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Spatial Reasoning on RefSpatial-Bench

72.2Overall Score

Seed

-2.57616.83736.2555.663Dec 9, 2025Dec 29, 2025Jan 18, 2026Feb 7, 2026Feb 27, 2026Mar 19, 2026Apr 8, 2026
Updated 8d ago

Evaluation Results

MethodLinks
2026.04
72.2--
2026.04
61--
2026.04
57.2--
2026.01
545454
2025.12
5151.9950
2025.12
50.7549.552
2026.04
43.3--
2026.01
42.544.541
2025.12
41.0746.136.03
38.1644.5831.73
2026.04
33.2--
2026.01
32.53629
30.2537.523
2026.01
19.6723.515.83
17.291519.58
2026.01
13.7216.8310.6
2025.12
11.2412.999.5
2026.01
9.9214.435.41
8.7889.55
7.69510.37
2025.12
6.9610.473.46
2026.01
5.449.841.04
2026.01
4.518
2026.01
0.30.030.57