GDPval-AA v2

economically valuable knowledge work (v2, re-based Elo)

Claude Opus 5 by Anthropic holds the best GDPval-AA v2 result among tracked models, at 1861 (released Jul 24 2026). Higher scores are better on this benchmark.

1. Claude Opus 5 (Anthropic) — 1861; 2. Claude Fable 5.1 (Anthropic) — 1853; 3. GLM-5.3-Flash (Z.ai) — 1773; 4. GLM-5.3 (Z.ai) — 1769; 5. Claude Fable 5 (Anthropic) — 1760.

20 tracked models have a published GDPval-AA v2 score. Scores are collected from published lab reports and benchmark sources, with recorded sources credited alongside the results. We do not re-run the evaluations.

GLM-5.3-Flash by Z.ai is the highest-ranked model with downloadable weights on GDPval-AA v2, scoring 1773 at rank 3 overall.