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

Scientific Code Generation on ScienceAgentBench (test)

32.8SR

MAA

9.9215.8621.827.74Mar 3, 2026Mar 20, 2026Apr 7, 2026Apr 25, 2026May 13, 2026May 31, 2026Jun 18, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
32.8--
2026.06
31.9--
2026.06
30.7--
2026.06
30.3--
2026.06
29.2--
2026.06
28.6--
2026.06
28.4--
2026.03
27.586.373.5
2026.03
27.586.487.3
2026.06
27.2--
2026.03
26.585.190.2
2026.06
26.1--
2026.06
25.8--
2026.03
23.585.671.6
2026.06
23.4--
2026.06
22.7--
2026.03
22.684.483.3
2026.06
22.6--
2026.06
22.1--
2026.06
21.7--
2026.06
20.8--
2026.06
20.4--
2026.06
19.8--
2026.03
19.683.178.4
2026.06
19.5--
2026.06
18.3--
2026.06
17.5--
2026.06
15.7--
2026.06
14.8--
2026.03
11.882.652.9
2026.06
11.3--
2026.03
10.883.841.2