GDPval

GDPval is a benchmark released by OpenAI in 2025 that evaluates AI models on real-world work tasks from 44 occupations by comparing their output with that of professionals.

2 articlesLast mentioned

GDPval is a benchmark released by OpenAI in September 2025 that measures how well AI models perform economically valuable real-world work. It covers 44 occupations drawn from nine industry sectors that contribute most to U.S. GDP, with tasks created by experienced professionals in each field. The deliverables take the form of documents, slide decks, spreadsheets, diagrams, and other artifacts used in actual work.

Evaluation is done by expert graders who compare a model's output with that of a human professional in a blind comparison and choose the better one. Rather than scoring problems with a single correct answer, the benchmark is intended to measure professional-quality work and is known as an attempt to gauge the economic impact of AI models.

This entry is based on AIPOST articles and widely known facts. If something is wrong, please send us a correction request.

Articles covering this entry

  • Anthropic S-1 RSI Disclosure: 2x Research Speedup, Not a Runaway LoopAI News ·

    He praises SemiAnalysis's GDPval and GDP-X benchmarks as a promising attempt to capture nuanced, realistic tasks in fields like finance, where ground truth is h…

    Read article
  • ChatGPT for Financial Services: One Shared Model, Hosted Data, Human ReviewAI News ·

    OpenAI's own GDPval benchmark, which spans 44 occupations, became saturated on general knowledge tasks and contained broken, uninformative items.

    Read article

Found a mistake or missing an entry?

Wiki entries are based on AIPOST articles and widely known facts. If an entry is wrong, send us a correction request and we will check and fix it. Names from new articles are added to the wiki after the article is published.