Which AI model scores highest on DROP?

o1-mini by OpenAI holds the best DROP result among tracked models, at 83.9 (released Sep 12 2024). Higher scores are better on this benchmark.

What are the top 5 models on DROP?

1. o1-mini (OpenAI) — 83.9; 2. GPT-4o (OpenAI) — 83.7; 3. GPT-4.5 (OpenAI) — 83.4; 4. GPT-4 Turbo (OpenAI) — 83.2; 5. GPT-4.1 nano (OpenAI) — 82.2.

How many models have a published DROP score?

15 tracked models have a published DROP 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.

Reasoning

DROP

three-shot token F1 on questions requiring reasoning over paragraphs

Rankings

Higher is better

DROP — frequently asked questions

: three-shot token F1 on questions requiring reasoning over paragraphs
: o1-mini by OpenAI holds the best DROP result among tracked models, at 83.9 (released Sep 12 2024). Higher scores are better on this benchmark.
: 1. o1-mini (OpenAI) — 83.9; 2. GPT-4o (OpenAI) — 83.7; 3. GPT-4.5 (OpenAI) — 83.4; 4. GPT-4 Turbo (OpenAI) — 83.2; 5. GPT-4.1 nano (OpenAI) — 82.2.
: 15 tracked models have a published DROP 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.