SemiAnalysis Releases GDPx Benchmark, No Model Exceeds 12%
Understand the story
On November 6, 2026, SemiAnalysis released the GDPx real-work benchmark. The benchmark places AI agents in simulated work environments lasting up to eight hours, comprising 206 tasks across 11 fields including chip design, biology, quantitative trading, cybersecurity, and law. According to reports, no AI agent has yet scored above 12% on the benchmark. No further developments have been reported.
AI overview of reporting · Updated 3 h前
Reporting timeline
Explore different aspects of the story through its reporting.
- The DecoderNo AI agent scores above 12 percent on SemiAnalysis' new real-world work benchmark
SemiAnalysis has released the GDPx benchmark, which places AI agents in simulated work environments of up to eight hours, with 206 tasks spanning 11 fields including chip design, biology, quant trading, cybersecurity and law.
Story heat trend
Current heat9·Comparable peak10（11/6 02:00）·Comparable change over 24 hours –
The trend compares the same continuously observed sources, which may cover fewer sources than the current heat score. Hover or select the chart to inspect each hour; use the left and right arrow keys to navigate.