SemiAnalysis launches GDPx, an 11-domain agentic benchmark where the best model scores 12%

SemiAnalysis published GDPx on Nov. 2, a benchmark of agentic tasks in 11 domains including chip design, cybersecurity, biology, finance, legal and software engineering. Headline score is pass@1 over three attempts, with an 8-hour default limit per task. SemiAnalysis says even the best model scores just 12% overall and calls Opus 5.5 the best all-around model.

HEAT

OUTLETS

1

NEW IN 6H

+1

FIRST SEEN

Nov 2, 2026, 00:40 ET

LAST SIGNAL

4h ago