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

Goal-conditioned visual planning on CrossTask T=4 88 (test)

37.04SR

GeoWorld ViT-g384

Updated 15d ago

Evaluation Results

Method	Links
GeoWorld ViT-g384 2026.02		37.04	71.35	87.04
GeoWorld ViT-g 2026.02		35.49	71	84.5
V-JEPA 2 ViT-g384 2026.02		35.01	70.24	85.05
GeoWorld ViT-H 2026.02		34.51	68.89	82.95
V-JEPA 2 ViT-g 2026.02		33.46	69.26	82.61
V-JEPA 2 ViT-H 2026.02		32.18	67.23	80.84
Gemini 2.5 Pro 2026.02		31.53	60.58	84.56
GeoWorld 2026.02		31.48	67.3	82.48
V-JEPA 2 2026.02		31.36	65.45	69.21
GeoWorld ViT-L 2026.02		30.63	65.46	79.73
GeoWorld 2026.02		30.63	66.02	81.82
GPT-5 2026.02		30.2	64.48	82.15
V-JEPA 2 2026.02		30.03	65.04	67.93
GeoWorld 2026.02		29.19	65.47	80.16
V-JEPA 2 ViT-L 2026.02		28.86	64.34	78.4
Qwen3-VL-Max 2026.02		28.76	62.91	81.51
V-JEPA 2 2026.02		28.75	64.71	66.82
InternVL3.5-241B 2026.02		27.65	63.54	80.13
GeoWorld 2026.02		27.64	64.35	79.43
V-JEPA 2 2026.02		27.53	63.8	65.45
VideoWorld 2026.02		25.5	60.26	76.85
MTID 2026.02		24.76	60.69	67.67
SCHEMA 2026.02		24.5	58.48	76.48
KEPP 2026.02		24.15	59.05	66.64
ActionDiffusion 2026.02		22.43	59.42	66.04
PDPP 2026.02		21.48	57.82	65.13
E3P 2026.02		16.49	48	70.16
Ext-MGAIL 2026.02		16.41	43.05	60.93
LFP 2026.02		15.97	50.7	75.3
VidAssist 2026.02		15.45	51.51	72.61
PlaTe 2026.02		14	35.29	55.36
P3IV 2026.02		13.4	44.16	70.01
VidAssist 2026.02		9.89	40.85	70.35
Int-MGAIL 2026.02		9.47	37.16	57.24
DDN 2026.02		5.97	27.1	48.46
Retrieval-Based 2026.02		3.95	22.22	36.97
UPN 2026.02		1.19	21.59	27.85
UAAA 2026.02		0.98	19.86	27.09
WLTDO 2026.02		0.77	17.92	26.43
Random 2026.02		0.01	1.83	1.66