aiqbench analysis
GDPx introduces a novel benchmarking approach that prioritizes high-ROI sectors like finance and hardware design over broad economic metrics, revealing how safety classifiers artificially suppress scores for frontier models in critical domains such as cybersecurity and biology. The results highlight distinct performance gaps between leading closed-source models and open-weight alternatives, while demonstrating that specific models excel in niche areas when evaluated without restrictive safety filters.
GDPx: Measuring Frontier Model Capabilities Across the Most Valuable Domains

From SemiAnalysis
ExcerptGDPx is a benchmark that covers the most valuable domains for frontier intelligence. It aims to be the world’s best benchmark for measuring general agentic capabilities.
We believe the first trillion dollars of AI revenue will be concentrated in sectors that can deploy frontier intelligence at high ROI, not broadly diffused throughout the whole economy. Building a general-purpose benchmark based on share of GDP or the SOC/O*NET occupational taxonomy is therefore a mistake. If your goal is to track the progress of frontier models, you should focus on the domains where they will be most impactful.
And intentionally forgo industries like government, real estate, and insurance.
GDPx is a true frontier-level benchmark: even the best model scores just 12%.
It’s worth emphasizing that certain models like Opus 5.5 and GPT-6 Astra have sandbagged overall scores due to their safety classifiers. If we exclude cyber, biology, and alignment—the three domains where safety classifiers sometimes kick in—you get the following results instead.
We were pleasantly surprised by how well this ranking lines up with our subjective impression of the models. OpenAI and Anthropic generally dominate the top of the ranking while SpaceXAI, Meta, and open-weight models are all meaningfully behind.
Excerpt shown under the publisher's RSS terms · 198 words · full text and figures are on the publisher's site
Related coverage

SOTA alignment assessments don’t strongly update us against misalignment
Redwood Research · Jul 31, 2026

Looping a transformer increases no-CoT capabilities more than general capabilities
AI Alignment Forum · Nov 12, 2026

How an assessor could investigate evidence of acceleration from the automation of AI R&D
METR · Oct 14, 2026

Evaluating Red Team and Blue Team Capability for AI Control Research
LessWrong · Jul 24, 2026