A reading-comprehension benchmark that requires discrete reasoning operations such as counting, comparison, and sorting.
Browse the latest scores, model modes, release dates, and parameter sizes for DROP.
Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology
| Rank | Model | License | |||
|---|---|---|---|---|---|
 Pangu Pro MoE Standard Mode | 91.20 | 2025-06-30 | 71.9B | Free Commercial | |
 ERNIE-4.5-300B-A47B Standard Mode | 91.10 | 2025-06-30 | 300B | Free Commercial | |
 DeepSeek-V3-0324 Standard Mode | 89.70 | 2025-03-24 | 671B | Free Commercial | |
4 |  GPT-4.1 Standard Mode | 89.20 | 2025-04-14 | Unknown | Closed |
5 |  Qwen3-235B-A22B Standard Mode | 88.70 | 2025-04-28 | 235B | Free Commercial |
6 |  Claude3-Opus Standard Mode | 83.10 | 2024-03-04 | Unknown | Closed |
7 | GPT-4 Standard Mode | 80.90 | 2023-03-14 | 175B | Closed |
8 |  Gemma 3 - 27B (IT) Standard Mode | 77.20 | 2025-03-12 | 27B | Free Commercial |
9 | Gemma2-27B Standard Mode | 74.20 | 2024-05-14 | 27B | Free Commercial |
