Mind2Web
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Mind2Web SC
95.8LPA
10
Multimodal-Mind2Web (out-of-distribution)
57.6Cross-Task Generalization
10
Mind2Web cross-Task
48.62Action Efficiency (EA)
9
Mind2Web OOD (test)
51.56Step Accuracy
9
Online-Mind2Web (Hard)
53.52Success Rate
9
Online-Mind2Web (Medium)
66.91Success Rate
9
Mind2Web-SC (test)
0.99LPA
9
Mind2Web (overall)
30.87Action Accuracy
8
Mind2Web cross-domain
26.97Action Accuracy
8
Multimodal-Mind2Web
66Step Success Rate
8
Mind2Web (Cross-Website)
46.1Execution Accuracy (EA)
8
Mind2Web Cross-domain
63.4Success Rate (Acc)
8
Mind2Web
92.44BS
8
Mind2Web
84Benign Utility
8
Mind2Web
18.86Task Success Rate
8
Mind2Web (test)
31.3Step-wise Success Rate
8
Multimodal-Mind2Web
49.1Element Accuracy
8
Multimodal-Mind2Web Cross-Website
48.1Element Accuracy
8
Mind2Web Cross-Domain GPT-4.1
63.4Success Rate (SR)
7
Mind2Web Cross-Website GPT-4.1
61.3Success Rate (SR)
7
Mind2Web Cross-Task GPT-4.1
63.2Success Rate
7
MIND2WEB Cross-Domain 1.0
42.1Element Accuracy
7
Mind2Web Cross-Web
6.91Average Steps per Successful Task
6
Mind2Web Cross-Task
7Avg Steps per Success
6
Mind2Web 2
30.7Navigation Score
6