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

Question Answering on CORTEXBENCH Bridge

96.1Accuracy

GPT-5.4

23.8242.58561.3580.115Jun 29, 2026
Updated 25d ago

Evaluation Results

MethodLinks
2026.06
96.1
2026.06
94.8
2026.06
91.9
2026.06
90.2
2026.06
84
2026.06
82.5
2026.06
80.1
2026.06
77.5
2026.06
65.7
2026.06
63
2026.06
62.5
2026.06
60.9
2026.06
59.1
2026.06
58.4
2026.06
54.1
2026.06
52.5
2026.06
52.5
2026.06
52.2
2026.06
50.9
2026.06
47.2
2026.06
44
2026.06
41.3
2026.06
38.1
2026.06
35.3
2026.06
34.7
2026.06
33.8
2026.06
30.9
2026.06
30.9
2026.06
30.3
2026.06
28.1
2026.06
26.9
2026.06
26.6