GDPVal-AA v2

GDPVal-AA v2 is a variant of GDPVal, OpenAI's benchmark of real, paid knowledge-work tasks from almost 50 professions (reports, analyses, presentations). Artifacts are judged against professional deliverables; the score is an index, not a percentage. It measures usefulness for actual work — not just exam-style questions.

GDPVal-AA v2 measures how well a model acts as a general-purpose agent. In detail, it checks:

• realistic professional work products
• deliverables that meet expert-defined quality criteria
• long tasks with documents, data and constraints
• breadth across many different occupations

The score is a composite index published by the benchmark organiser rather than a plain percentage. Interpret it only against the other models in the table — higher is better.

What is the GDPVal-AA v2 benchmark?

GDPVal-AA v2 is a variant of GDPVal, OpenAI's benchmark of real, paid knowledge-work tasks from almost 50 professions (reports, analyses, presentations). Artifacts are judged against professional deliverables; the score is an index, not a percentage. It measures usefulness for actual work — not just exam-style questions.

What does GDPVal-AA v2 test?

GDPVal-AA v2 mainly measures how well a model acts as a general-purpose agent. Concretely, it checks: realistic professional work products; deliverables that meet expert-defined quality criteria; long tasks with documents, data and constraints.

How is the GDPVal-AA v2 score calculated?

The score is a composite index published by the benchmark organiser rather than a plain percentage. Interpret it only against the other models in the table — higher is better.

Where do I get GDPVal-AA v2 data and scores?

The official benchmark page/reference is linked on this page. Every score on Open LLM Radar is also available as open data: download the scores JSON (http://open-llm-radar-api:8000/api/v1/benchmarks/gdpval-aa-v2/scores) or pull it directly from the Open LLM Radar API with any HTTP client.