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

Interactive Environment Task Completion on ScienceWorld (Unseen)

90.1Average Reward

EAGLET + GPT-5

-0.3823.1146.670.09Oct 7, 2025Nov 16, 2025Dec 26, 2025Feb 4, 2026Mar 16, 2026Apr 25, 2026Jun 5, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2025.10
90.1
2025.10
89
2025.10
88.2
2025.10
85.6
2025.10
83.8
2025.10
83.6
2025.10
82.5
2025.10
80.8
2025.10
79.9
2025.10
78.2
2025.10
76.5
2025.10
76.2
2025.10
74.5
2025.10
74.1
2026.06
69.7
2025.10
69.6
2025.10
68.4
2026.06
66.4
2025.10
65.9
2026.06
65
2026.06
64.4
2026.06
64.4
2026.06
61.7
2025.10
61.6
2025.10
58.1
2026.06
57.6
2025.10
57
2025.10
55.5
2026.06
54.3
2026.06
53
2026.06
51.7
2025.10
42.2
2026.06
13
2026.06
3.1