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

Visual Planning on CrossTask

38.45Success Rate (SR)

ViterbiPlanNet

Updated 1mo ago

Evaluation Results

Method	Links
ViterbiPlanNet 2026.03		38.45	63.07	83.89
SCHEMA 2026.03		37.24	62.69	83.94
PlanLLM 2026.03		36.84	61.56	83.23
PDPP 2026.03		36.73	61.96	83.2
KEPP 2026.03		34.93	60.34	82.67
Gemini 2.5 Pro 2026.03		29.18	57.9	81.48
Qwen2.5-32B 2026.03		25.14	56.1	80.92
ViterbiPlanNet 2026.03		24.64	57	81.18
SCHEMA 2026.03		24.18	57.02	81.46
Qwen3-30B 2026.03		23.37	55.96	81.16
Qwen3-30B + PKG 2026.03		23.31	56.15	81.06
PlanLLM 2026.03		22.91	55.29	81.03
PKG beam search 2026.03		22.38	55.74	80.92
KEPP 2026.03		22.34	55.24	80.58
PDPP 2026.03		21.47	55.66	80.68
Gemini 2.5 Pro 2026.03		14	51.33	78.58
Qwen2.5-VL-32B 2026.03		11.48	36.35	69.52
Qwen3-30B + PKG 2026.03		10.96	48.77	77.48
Qwen3-30B 2026.03		10.59	49.06	78.03
PKG beam search 2026.03		9.3	47.65	78.25
Qwen2.5-32B 2026.03		9.22	46.32	76.15
Qwen2.5-VL-32B 2026.03		5.56	31.22	66.31