OpenAI evaluation spanning well-specified knowledge work across 44 occupations. Measures whether AI matches or exceeds industry professionals.
Broad coverage across 44 professional occupations; practical knowledge-work tasks rather than academic benchmarks.
Self-reported by OpenAI; methodology details limited; may not generalize to all professional contexts.
Higher is better. Percentage of comparisons where AI matches or exceeds professionals.
Max score: 100
Self-reported by OpenAI.
| # | Model | Score |
|---|---|---|
| 1 | 83.0 |