STEMVerse - Mathematics - Differential Equations and Dynamical Systems: leaderboard
STEMVerse mathematics problems labelled with the sub-discipline 'Differential Equations and Dynamical Systems', drawn from existing STEM benchmarks pooled into one set; labels come from GPT-4o and were checked by an expert audit. Few-shot prompting in the MegaScience style; accuracy in percent, higher is better. Li et al. (ZGCA and NTU), 2026 preprint.
Metric: Few-shot accuracy (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 16 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen3-8B-Base | 59.09 | #1728 |
| 2 | Qwen3-14B-Base | 54.99 | #1608 |
| 3 | Qwen3-4B-Base | 53.52 | #2444 |
| 4 | Qwen 2.5 3B Instruct | 51.91 | #5206 |
| 5 | Qwen 2.5 7B | 50.44 | #2718 |
| 6 | Qwen 2.5 14B | 49.56 | #1794 |
| 7 | Qwen 2.5 7B Instruct | 49.41 | #2583 |
| 8 | Qwen 2.5 14B Instruct | 42.23 | #1745 |
| 9 | Llama 3.1 8B Instruct | 41.94 | #4277 |
| 10 | Llama 3.2 3B Instruct | 36.51 | #7185 |
| 11 | Qwen 2.5 3B | 36.22 | #5214 |
| 12 | Llama 3.1 8B | 21.85 | #4900 |
| 13 | Llama 3.2 3B | 9.24 | #7877 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=stemverse-mathematics-differential-equations-and-dynamical-systems · How It Works · Data refreshed daily, snapshot 2026-11-12.