Equational Theories - Normal (Low-or-none reasoning): leaderboard
Metric: Strict F1 (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 25 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3.5 122B A10B | 97.53 |
| 2 | GPT-5.4 | 97.19 |
| 3 | Gemini 3.1 Pro (Preview) | 96.23 |
| 4 | Qwen 3.5 397B A17B | 91.03 |
| 5 | Claude Opus 4.6 | 84.75 |
| 6 | Gemini 3 Flash (Preview) | 84.23 |
| 7 | Claude Sonnet 4.6 | 82.8 |
| 8 | GPT-5 Mini | 81.07 |
| 9 | Seed 2.0 Lite | 77.4 |
| 10 | Kimi K2.5 | 74.03 |
| 11 | GPT-OSS-120B | 71.74 |
| 12 | Qwen 3.5 27B | 68.22 |
| 13 | GLM-5 | 61.94 |
| 14 | MiniMax-M2.5 | 60.34 |
| 15 | GPT-5 Nano | 51.98 |
Interactive version: theaggregate.ai/benchmark?slug=equational-theories-normal-low-or-none-reasoning · How It Works · Data refreshed daily, snapshot 2026-10-09.