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

Science Question Answering on GPQA 0-shot (test)

162.95Throughput

d2Cache

-0.527641.913784.355126.7963Sep 27, 2025Oct 25, 2025Nov 23, 2025Dec 22, 2025Jan 20, 2026Feb 18, 2026Mar 19, 2026
Updated 29d ago

Evaluation Results

MethodLinks
2025.09
162.9525.32.0832.8
2025.09
89.0314.56.8928.8
2026.03
38.73--26.79
2026.03
35.23--27.01
2026.03
34.69--25.22
2026.03
26.9--27.68
2026.03
25.34--24.78
2026.03
23.22--22.1
2026.03
20.98--24.55
2026.03
19.12--24.78
2026.03
18.08--24.78
2026.03
11.22--27.68
2026.03
6.14--27.01
2026.03
5.76--27.68