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

General AI Assistant Task on GAIA (accuracy, delta)

84.5Accuracy

Step-3.5-Flash

0.98822.66944.3566.031May 16, 2026May 19, 2026May 23, 2026May 27, 2026May 31, 2026Jun 4, 2026Jun 8, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
84.5-
2026.06
82.5-
2026.06
80.3-
2026.06
80.1-
2026.06
78.6-
2026.06
77.7-
2026.06
76.4-
2026.06
75.7-
2026.06
75.1-
2026.06
74.8-
2026.06
72-
2026.06
71.5-
2026.06
70.9-
2026.06
70.9-
2026.06
66-
2026.05
62-
2026.05
55.7-
2026.05
55.7-
2026.05
55-
2026.05
52.1-
2026.05
51.3-
2026.05
50.4-
2026.05
47.312.3
2026.05
45-
2026.05
38.610.4
2026.05
35-
2026.05
35-
2026.05
33.3-
2026.05
28.76.7
2026.05
28.2-
2026.05
28.2-
2026.05
22-
2026.05
16.18.1
2026.05
8-
2026.05
6.4-
2026.05
4.2-