Discrete Reasoning Over Paragraphs: reading comprehension with discrete reasoning (numbers, dates, counts) over passages.
Models ranked by score on this benchmark. Higher is better.
| 3 |
| Anthropic |
| 86.5 |
| p96 |
| Code Assistant, Small, Text Generation, Multimodal, Reasoning, Proprietary |
| 4 | 85.8 | p95 | Multimodal, Small, Text Generation, Reasoning, Proprietary |
| 5 | Anthropic | 85.5 | p95 | Multimodal, Small, Text Generation, Reasoning, Proprietary, Large |
| 6 | DeepSeek | 85.0 | p95 | Text Generation, Reasoning, Open Weight, Large |
| 7 | Alibaba | 84.5 | p94 | Reasoning, Large, Code Assistant, Text Generation, Multimodal, Open Weight |
| 8 | Alibaba | 82.5 | p91 | Text Generation, Large, Multimodal, Reasoning, Open Weight |
| 9 | Meta | 82.0 | p92 | Reasoning, Large, Text Generation, Open Weight |
| 10 | Meta | 81.5 | p91 | Text Generation, Reasoning, Open Weight, Medium |
| 11 | Mistral AI | 80.5 | p91 | Text Generation, Small, Reasoning, Proprietary |
| 12 | Alibaba | 79.0 | p88 | Code Assistant, Medium, Reasoning, Text Generation, Open Weight |
| 13 | 77.0 | p86 | Reasoning, Medium, Code Assistant, Text Generation, Open Weight |
| 14 | OpenAI | 76.8 | p87 | Code Assistant, Small, Text Generation, Multimodal, Proprietary |
| 15 | Alibaba | 76.0 | p86 | Code Assistant, Small, Text Generation, Reasoning, Open Weight, Medium |
| 16 | Anthropic | 72.5 | p83 | Multimodal, Small, Text Generation, Proprietary |
| 17 | Microsoft | 70.0 | p79 | Multimodal, Medium, Text Generation, Open Source, Small |
