STEMVerse - Physics - Astrophysics: leaderboard
Metric: Few-shot accuracy (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 16 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen3-14B-Base | 39 |
| 2 | Qwen 2.5 14B | 33 |
| 3 | Qwen 2.5 14B Instruct | 31 |
| 4 | Qwen3-8B-Base | 30 |
| 5 | Qwen 2.5 7B Instruct | 28 |
| 6 | Qwen 2.5 7B | 27 |
| 7 | Llama 3.1 8B Instruct | 24 |
| 8 | Qwen 2.5 3B Instruct | 22.58 |
| 9 | Llama 3.2 3B Instruct | 21 |
| 10 | Qwen 2.5 3B | 18 |
| 11 | Llama 3.2 3B | 17 |
| 12 | Llama 3.1 8B | 16 |
| 13 | Qwen3-4B-Base | 13 |
Interactive version: theaggregate.ai/benchmark?slug=stemverse-physics-astrophysics · How It Works · Data refreshed daily, snapshot 2026-09-25.