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

Goal-conditioned visual planning on COIN T=4 71 (test)

33.29Success Rate (SR)

GeoWorld ViT-g384

Updated 15d ago

Evaluation Results

Method	Links
GeoWorld ViT-g384 2026.02		33.29	61.56	91.84
GPT-5 2026.02		32.64	56.84	86.38
V-JEPA 2 ViT-g384 2026.02		31.63	59.28	90.51
Gemini 2.5 Pro 2026.02		30.2	60.13	84.82
GeoWorld ViT-g 2026.02		29.93	60.65	91.81
V-JEPA 2 ViT-g 2026.02		29.6	57.73	90.23
GeoWorld ViT-H 2026.02		28.82	58.1	90.48
V-JEPA 2 ViT-H 2026.02		27.38	56.07	89.2
GeoWorld ViT-L 2026.02		26.4	54.52	86.98
Qwen3-VL-Max 2026.02		26.17	57.13	87.56
InternVL3.5-241B 2026.02		25.46	55.3	88.22
V-JEPA 2 ViT-L 2026.02		25.29	53.3	86.21
VideoWorld 2026.02		23.74	51.27	85.33