DROP — DROP with explicit 3-shot F1 metric; matches english_drop_3shot_f1 column
| # | Model | Vendor | Best score | Runs | Last seen |
|---|---|---|---|---|---|
| 1 | DeepSeek-R1 | DeepSeek | 92.2 | 6 | 2026-06-04 |
| 2 | DeepSeek-V3 | DeepSeek | 91.6 | 6 | 2026-06-04 |
| 3 | o1 | OpenAI | 90.2 | 5 | 2026-06-04 |
| 4 | R1-Zero | DeepSeek | 89.1 | 1 | 2026-05-03 |
| 5 | Llama 3.1 Instruct 405B | Meta | 88.7 | 1 | 2026-05-03 |
| 6 | Claude 3.5 Sonnet | Anthropic | 88.3 | 11 | 2026-06-05 |
| 7 | DeepSeek-V2.5 | DeepSeek | 87.8 | 1 | 2026-05-03 |
| 8 | R1-Distill-Qwen-14B | Alibaba | 85.5 | 5 | 2026-06-05 |
| 9 | OpenAI o1-mini | OpenAI | 83.9 | 10 | 2026-06-05 |
| 10 | GPT-4o | OpenAI | 83.7 | 11 | 2026-06-05 |
| 11 | DeepSeek-V2 | DeepSeek | 83 | 1 | 2026-05-03 |
| 12 | MiMo 7B RL | Xiaomi | 78.7 | 5 | 2026-06-05 |
| 13 | R1-Distill-Qwen-7B | Alibaba | 77 | 5 | 2026-06-05 |
| 14 | Qwen2.5 Instruct 72B | Alibaba | 76.7 | 1 | 2026-05-03 |
| 15 | QwQ 32B Preview | Alibaba | 71.2 | 5 | 2026-06-05 |
Best tracked score per model (default configuration; source-attributed and verification-tiered). Open a model for its full benchmark surface, provenance and pricing.