Artificial Intelligence1 min read

SemiAnalysis launches GDPx, an agentic capability benchmark measuring 11 domains, including chip design, cybersecurity, and RSI; Opus 5.5 tops with a 12% score (SemiAnalysis)

By · Published by Everything Blog

In short

Designed to measure agentic capabilities across 11 diverse areas, the assessment tests systems on intricate tasks such as materials science and recursive self-improvement. The benchmark also encompasses fields like software engineering, quantitative trading, and AI alignment. The leading system, Opus 5.5, secured first place overall while obtaining a score of only 12 percent.

Key points

  • Leading artificial intelligence models still struggle significantly with end-to-end tasks…: Leading artificial intelligence models still struggle significantly with end-to-end tasks across highly specialized industries.
  • Agentic evaluations are prioritizing rigorous real-world domains such as chip design, fin…: Agentic evaluations are prioritizing rigorous real-world domains such as chip design, finance, and cybersecurity.
  • Current frontier capabilities leave a substantial margin for improvement across complex t…: Current frontier capabilities leave a substantial margin for improvement across complex technical problem-solving benchmarks.

Sources checked by our analysis pipeline: 1

SemiAnalysis : SemiAnalysis launches GDPx, an agentic capability benchmark measuring 11 domains, including chip design, cybersecurity, and RSI; Opus 5.5 tops with a 12% score   —  A comprehensive benchmark spanning RSI, hardware design, cyber, bio, finance, legal, quant trading, chip design, software engineering, materials science, and alignment

Artificial Intelligence