GDPx: Measuring Frontier Model Capabilities Across the Most Valuable Domains

SemiAnalysis has introduced GDPx, a new benchmark designed to measure general agentic capabilities across the most valuable domains for frontier intelligence. The benchmark was developed in collaboration with specialized RL environment startups to evaluate models on complex, realistic environments rather than static, one-shot prompts. The authors argue that this approach addresses the shortcomings of older benchmarks like OpenAI's GDPval by avoiding broken rubrics and lack of domain-specific workflows. Open article →