STEMVerse - Mathematics - Algebra and Geometry: leaderboard
Metric: Few-shot accuracy (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 16 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen3-8B-Base | 55.09 |
| 2 | Qwen3-14B-Base | 52.06 |
| 3 | Qwen 2.5 3B Instruct | 50.94 |
| 4 | Qwen 2.5 14B | 49.88 |
| 5 | Qwen3-4B-Base | 48.42 |
| 6 | Qwen 2.5 7B | 46.51 |
| 7 | Qwen 2.5 7B Instruct | 40.25 |
| 8 | Qwen 2.5 14B Instruct | 38.62 |
| 9 | Qwen 2.5 3B | 37.67 |
| 10 | Llama 3.1 8B Instruct | 36.34 |
| 11 | Llama 3.2 3B Instruct | 30.79 |
| 12 | Llama 3.1 8B | 20.38 |
| 13 | Llama 3.2 3B | 8.74 |
Interactive version: theaggregate.ai/benchmark?slug=stemverse-mathematics-algebra-and-geometry · How It Works · Data refreshed daily, snapshot 2026-09-25.