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

Interactive Agent Task Completion on ScienceWorld

90Average Reward

OpenSkill

75.33679.14382.9586.757Jun 4, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
90
2026.06
88.7
2026.06
88.3
2026.06
88
2026.06
88
2026.06
87.1
2026.06
85.3
2026.06
83.1
2026.06
80
2026.06
78
2026.06
77.8
2026.06
75.9