WinoGrande

resolving ambiguous pronouns in context

GPT-4 by OpenAI holds the best WinoGrande result among tracked models, at 87.5% (released Mar 14 2023). Higher scores are better on this benchmark.

1. GPT-4 (OpenAI) — 87.5%; 2. LLaMA 2 (Meta) — 80.2%; 3. LLaMA 1 (Meta) — 77%.

3 tracked models have a published WinoGrande 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.