AGC-Bench - puntuguese — leaderboard
Metric: Dataset z-score. Source: huggingface.co. 83 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | granite-4.0-h-micro | 2.49 |
| 2 | Nova Lite (v1) | 1.48 |
| 3 | Mistral Nemo | 1.32 |
| 4 | Llama 3.2 11B Instruct | 1.02 |
| 5 | Llama 3.1 8B Instruct | 0.97 |
| 6 | Gemini 2.5 Flash | 0.94 |
| 7 | Claude Sonnet 4 | 0.94 |
| 8 | Gemma 4 31B (IT) | 0.94 |
| 9 | DeepSeek V3.1 Terminus | 0.94 |
| 10 | GPT-5.1 | 0.88 |
| 11 | GPT-5.4 Mini | 0.76 |
| 12 | Gemma 4 26B A4B (IT) | 0.76 |
| 13 | GPT-4.1 Mini | 0.7 |
| 14 | GPT-4.1 | 0.68 |
| 15 | GPT-5.4 | 0.67 |
Interactive version: theaggregate.ai/benchmark?slug=agc-bench-puntuguese · How the rankings work · Data refreshed daily, snapshot 2026-07-22.