Edition
No AI agent scores above 12 percent on SemiAnalysis' new real-world work benchmark
GDPx evaluates AI agents in simulated work settings; no model scores above 12 percent.
GDPx covers 206 tasks across eleven fields, from chip design and biology to cybersecurity and law, while excluding administration and insurance. In some fields, including law and hardware design, language models grade results using rubrics rather than human experts, leaving outsiders unable to fully verify task difficulty or grading quality.
This summary is based on the available full source text. Summary prepared with AI assistance. The original source is authoritative.
Original source
The Decoder
Date stated with the source: 5. 11. 2026
Open the original at the publisher ↗