← Back
software
GDPX
Last discussed Nov 4, 2026
SemiAnalysis benchmark measuring frontier AI model capability in high-value domains such as chip design, biology, and cybersecurity.
Recently discussed in
Big Tech’s $3T AI Bet, Claude Writes 90% of Anthropic’s Code, AIs Editing AI Brains | EP #308
Alexander Wissner-Gross · Nov 4, 2026
$AAPL + $PLTR, GDPx, Value of a Search, Grok Bot Breaks the Bank
Max Kan · Nov 3, 2026
Anthropic Q3 Revenue, Europe’s Freezer Hack Scare, Northwood Raises $375M | Diet TBPN
John Coogan · Nov 3, 2026
What creators say
AW
Alexander Wissner-Gross
mentioned
Praised: Best models score in the 10-20% sweet spot; modeled on frontier fields rather...
vs: Contrasted with Anthropic and OpenAI evals modeled on human occupations
"So many things to love about this benchmark. First of all, we're starved for good evals... GDPX is modeled after actually the frontier, like recursive self-improvement and material science."
More from Alexander Wissner-Gross →
John Coogan
mentioned
Praised: Focuses on practical, valuable business tasks
vs: Contrasted with mathy evals
"They have a big benchmark out, GDPX... GDPX is a benchmark for measuring agentic capabilities across the most valuable domains. And they highlight the fact that the best model scores just 12%."
More from John Coogan →Created by
MK
Max Kan
creator
Used for: Measuring frontier model capability
Praised: Focuses on domains likely to drive the first trillion dollars of AI revenue
vs: Contrasted with OpenAI's GDPval
"GDPX is, in our opinion, the single best benchmark for measuring frontier model capabilities... We made a benchmark focused exclusively on those domains, which are things like chip design, biology, cybersecurity, RSI, etc."
More from Max Kan →Keep exploring
Topics
About this data
- ✓48,000+ mentions extracted from real podcast transcripts
- ✓Every quote links to the source episode
- ✓Sponsored mentions are always labeled
- ✓Critical takes are shown, never filtered out
Get an email when a new creator recommends it