Language Generation on ALBA
51.1ScoreGemma 3-12B
Evaluation results
| Method | Links | |
|---|---|---|
| Gemma 3-12BModel Type=Open weight models2026.03 | 51.1 | |
| Qwen 3-8BModel Type=Open weight models2026.03 | 49.8 | |
| AMALIA-9B-DPOModel Type=Fully open models, Training Strategy=DPO2026.03 | 43.6 | |
| Gemma 2-9BModel Type=Open weight models2026.03 | 41.1 | |
| Apertus-8BModel Type=Fully open models2026.03 | 38.7 | |
| EuroLLM-9BModel Type=Fully open models2026.03 | 38.5 | |
| AMALIA-9B-SFTModel Type=Fully open models, Training Strategy=SFT2026.03 | 37.1 | |
| Ministral-8BModel Type=Open weight models2026.03 | 35.6 | |
| Llama 3.1-8BModel Type=Open weight models2026.03 | 31.3 | |
| Gervasio-8BModel Type=Open weight models2026.03 | 31.1 | |
| Qwen 2.5-7BModel Type=Open weight models2026.03 | 31 | |
| Salamandra-7BModel Type=Fully open models2026.03 | 27.4 | |
| Mistral-7BModel Type=Open weight models2026.03 | 21.7 | |
| OLMo 2-7BModel Type=Fully open models2026.03 | 16.9 |