AGC-Bench - proparalogy: leaderboard
Metric: Dataset z-score. Source: huggingface.co. 82 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Hermes 4 405B | 0.94 |
| 2 | GPT-5.5 | 0.9 |
| 3 | Olmo 3.1 32B Instruct | 0.87 |
| 4 | pixtral-large-2411 | 0.87 |
| 5 | Kimi K2.5 | 0.86 |
| 6 | Claude Sonnet 4.6 | 0.86 |
| 7 | Claude Sonnet 4.5 | 0.86 |
| 8 | GPT-5.4 | 0.82 |
| 9 | Qwen Turbo | 0.79 |
| 10 | GPT-5.1 | 0.78 |
| 11 | GPT-4.1 | 0.78 |
| 12 | Nova Lite (v1) | 0.78 |
| 13 | Mistral Medium 3.1 | 0.77 |
| 14 | GLM-4.6 | 0.71 |
| 15 | Gemini 2.5 Flash | 0.7 |
Interactive version: theaggregate.ai/benchmark?slug=agc-bench-proparalogy · How It Works · Data refreshed daily, snapshot 2026-09-05.