GDPx: Measuring Frontier Model Capabilities Across the Most Valuable Domains
SemiAnalysis has introduced GDPx, a benchmark designed to measure AI capabilities across 11 agentic domains. Claude Opus 5.5 currently leads the benchmark with a score of 12%, which represents the top performance across all tested models. The evaluation spans high-value technical fields including hardware and chip design, cyber, bio, finance, legal, quant trading, software engineering, materials science, alignment, and RSI.
Listen to Live Briefing
Real-time synthesized voice briefing · Live Feeds Desk
- ✓ SemiAnalysis launched the GDPx benchmark covering 11 agentic domains.
- ✓ The GDPx benchmark includes domains such as hardware design, chip design, cyber, bio, finance, legal, quant trading, software engineering, materials science, alignment, and RSI.
- ✓ Claude Opus 5.5 leads the GDPx benchmark with a score of 12%.
- ✓ The highest score achieved on the GDPx benchmark is 12%.
What changed
SemiAnalysis released the GDPx benchmark and identified Claude Opus 5.5 as the top-performing model.
Live updates
-
SemiAnalysis Launches GDPx Benchmark for Frontier AI Models
SemiAnalysis has introduced GDPx, a benchmark designed to measure AI capabilities across 11 agentic domains. Claude Opus 5.5 currently leads the benchmark with a score of 12%, which represents the top performance across all tested models. The evaluation spans high-value technical fields including hardware and chip design, cyber, bio, finance, legal, quant trading, software engineering, materials science, alignment, and RSI.
Why it matters
Existing benchmarks often fail to capture complex, multi-step reasoning in professional domains. GDPx targets these gaps by testing agentic capabilities in specialized industries. The low top score suggests a significant performance gap between current frontier models and expert-level proficiency.
What is confirmed
- SemiAnalysis launched the GDPx benchmark covering 11 agentic domains.
- The GDPx benchmark includes domains such as hardware design, chip design, cyber, bio, finance, legal, quant trading, software engineering, materials science, alignment, and RSI.
- Claude Opus 5.5 leads the GDPx benchmark with a score of 12%.
- The highest score achieved on the GDPx benchmark is 12%.
Still unconfirmed
- Anthropic's pre-IPO filing indicates Claude writes over 90% of its own code.
What to watch next
- Release of GDPx scores for other frontier models
- Updates to the benchmark as models evolve
- Publication of the full GDPx methodology by SemiAnalysis
confidence 100%Sources used for this update (13)
- www.cbssports.com — NBA News, Scores, Stats, Standings and Rumors - National Basketball Association
- www.nba.com — NBA.com - The official site of the NBA for the latest NBA Scores,...
- www.espn.com — NBA on ESPN - Scores, Stats and Highlights
- www.espn.com — NBA Scores, 2026-27 Season - ESPN
- newsletter.semianalysis.com — GDPx: Measuring Frontier Model Capabilities Across the Most Valuable Domains
- SemiAnalysis — GDPx: Measuring Frontier Model Capabilities Across the Most Valuable Domains
- Unite.AI — SemiAnalysis Launches GDPx Benchmark Spanning 11 Agentic Domains
- FourWeekMBA — Opus 5.5 Leads SemiAnalysis’s GDPx Agent Benchmark at 12%
- shiftdelete.net — GDPx AI Benchmark Tops Out at 12% as Claude Opus 5.5 Leads
- Crypto Briefing — SemiAnalysis unveils GDPx benchmark, and the best AI model scores 12%
- finance.biggo.com — Anthropic's Claude Writes 90% of Its Code as Five Tech Giants Commit $3 Trillion to AI
- en.wikipedia.org — National Basketball Association - Wikipedia
Community Sentiment: How do you assess this situation?
Voice your perspective · Real-time aggregated sentiment from the Live Feeds community