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

Terminal Agentic Trajectory Generation on TerminalBench 2.0

57.8Score

Claude-Opus-4.5

-2.31213.29428.944.506Jan 6, 2026Jan 10, 2026Jan 14, 2026Jan 19, 2026Jan 23, 2026Jan 27, 2026Feb 1, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.02
57.8
2026.02
56.9
2026.01
54.2
2026.02
54
2026.02
51.7
2026.02
47.6
2026.02
46.4
2026.01
46.4
2026.02
42.8
2026.01
42.8
2026.02
41
2026.01
38.5
2026.01
35.7
2026.01
35.2
2026.02
31.8
2026.02
29.2
2026.02
27.8
2026.02
23.9
2026.02
22
2026.02
19.1
2026.02
16.7
2026.02
14.6
2026.02
13.5
2026.02
10.1
2026.02
8.3
2026.02
4.49
2026.02
3.4
2026.02
1.18
2026.02
0