QuantEval - Quantitative Reasoning (CoT): leaderboard
Metric: Accuracy (%; chain-of-thought prompting). Source: arxiv.org. Saturation forecast: Around 2029. 13 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 | 55 |
| 2 | Claude Sonnet 4.5 | 43.1 |
| 3 | Qwen 3 30B A3B | 40.5 |
| 4 | Gemini 2.5 Pro | 38 |
| 5 | Qwen 3 14B | 35.4 |
| 6 | Qwen 3 8B | 33.9 |
| 7 | DeepSeek R1 Distill Qwen 1.5B | 30 |
| 8 | DeepSeek R1 Distill Qwen 14B | 29.2 |
| 9 | Qwen 3 4B | 27.7 |
| 10 | DeepSeek-R1-Distill-Qwen-7B | 21.5 |
| 11 | DeepSeek R1 Distill Llama 8B | 10.8 |
Interactive version: theaggregate.ai/benchmark?slug=quanteval-quantitative-reasoning-cot · How It Works · Data refreshed daily, snapshot 2026-09-25.