# GDPval

> OpenAI benchmark of real-world, economically valuable knowledge-work tasks across 44 occupations; a 220-task gold subset with files and rubrics is open on Hugging Face.

- id: `gdpval`
- category: Information (`information`)
- url: https://openai.com/index/gdpval/
- docs: https://huggingface.co/datasets/openai/gdpval
- status: active
- pricing: free
- tags: benchmark, knowledge-work, dataset, evals, economics
- related: TheAgentCompany (https://indexagentica.com/entries/the-agent-company/); Humanity's Last Exam (https://indexagentica.com/entries/humanitys-last-exam/); METR Task-Completion Time Horizons (https://indexagentica.com/entries/metr-time-horizons/)
- sources: https://huggingface.co/datasets/openai/gdpval , https://arxiv.org/abs/2510.04374 , https://openai.com/index/gdpval/
- added: 2026-10-26
- updated: 2026-10-26
- last verified: 2026-10-26
- maintainer: OpenAI
- submitted by: agentica-curator
- page: https://indexagentica.com/entries/gdpval/
- json: https://indexagentica.com/api/entries/gdpval.json

The full set has 1,320 tasks built from real work products by professionals averaging 14 years of experience, spanning 44 occupations in the top 9 sectors of U.S. GDP. Each task has a prompt plus reference files (spreadsheets, documents, etc.); the open gold subset (`openai/gdpval`, v2 with rubrics and gold deliverables) has 220 tasks. OpenAI offers an experimental automated grader at evals.openai.com; human expert comparison remains the recommended grading method. Some tasks contain NSFW or political content. The Hugging Face dataset card does not state a license.