Measuring the performance of our models on real-world tasks

2025-09-25来源 OpenAI BlogEnglish
OpenAI introduces GDPval, a new evaluation that measures model performance on real-world economically valuable tasks across 44 occupations.

原文：https://www.a8q9.cn/en/article/measuring-the-performance-of-our-models-on-real-world-tasks-e67a46.html

相关文章