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

Graduate-level Science Question Answering on GPQA (Accuracy & Loss)

61.95Accuracy (GPQA)

Zero-Shot CoT

18.301229.633140.96552.2969Mar 7, 2025May 19, 2025Jul 31, 2025Oct 12, 2025Dec 24, 2025Mar 7, 2026May 19, 2026
Updated 1d ago

Evaluation Results

MethodLinks
2025.03
61.95-
2025.10
61.4-
2025.10
60.1-
2025.10
60.1-
2025.10
59.5-
2025.03
57.78-
2025.10
56.3-
2026.05
54.51-
2025.10
53.8-
2025.10
53.8-
2025.10
53.8-
2025.03
53.54-
2025.10
53.2-
2025.10
53.2-
2026.05
53.16-
2025.03
52.86-
2025.03
52.86-
2026.05
52.73-
2026.05
52.72-
2026.05
52.54-
2025.10
52.5-
2025.03
51.44-
2025.10
51.3-
2025.10
51.3-
2025.03
51.02-
2026.05
51.01-
2025.03
50.84-
2025.10
50.6-
2025.10
50.6-
2025.10
50.6-
2025.03
50.51-
2026.05
50.45-
2026.05
50.41-
2025.10
50-
2025.10
50-
2025.10
50-
2025.10
50-
2026.05
49.72-
2025.10
49.6-
2026.05
49.33-
2025.03
48.43-
2026.05
48.32-
2025.10
48.1-
2025.03
47.98-
2025.10
47.5-
2026.05
47.48-
2026.05
47.21-
2026.05
47.2-
2025.10
46.8-
2025.10
46.8-
2025.10
46.8-
2025.10
46.8-
2025.10
46.2-
2026.05
45.96-
2025.03
44.95-
2025.10
42.4-
2025.10
39.9-
2026.05
33.3-
2026.05
33.1-
2026.05
33-
2026.04
32.82.005
2026.05
32-
2026.05
31.5-
2026.04
30.42.201
2026.04
29.61.981
2026.05
28.91-
2026.05
28.75-
2026.05
27.88-
2026.05
26.21-
2026.05
25.48-
2026.04
252.221
2026.05
19.98-