Portuguese Educational Proficiency on PT-E
82.8AccuracyQwen 3-8B
Evaluation results
| Method | Links | |
|---|---|---|
| Qwen 3-8BModel Category=Open weight models, Instruction-tuned=true, Reasoning=Chain-of-thought (CoT)2026.03 | 82.8 | |
| Gemma 3-12BModel Category=Open weight models, Instruction-tuned=true, Reasoning=Chain-of-thought (CoT)2026.03 | 81.3 | |
| Qwen 2.5-7BModel Category=Open weight models, Instruction-tuned=true, Reasoning=Chain-of-thought (CoT)2026.03 | 73.7 | |
| Gemma 2-9BModel Category=Open weight models, Instruction-tuned=true, Reasoning=Chain-of-thought (CoT)2026.03 | 70.9 | |
| AMALIA-9B-DPOModel Category=Fully open models, Training=DPO, Instruction-tuned=true, Reasoning=Chain-of-thought (CoT)2026.03 | 68.4 | |
| Llama 3.1-8BModel Category=Open weight models, Instruction-tuned=true, Reasoning=Chain-of-thought (CoT)2026.03 | 66.7 | |
| Gervasio-8BModel Category=Open weight models, Instruction-tuned=true, Reasoning=Chain-of-thought (CoT)2026.03 | 66 | |
| Ministral-8BModel Category=Open weight models, Instruction-tuned=true, Reasoning=Chain-of-thought (CoT)2026.03 | 64.9 | |
| AMALIA-9B-SFTModel Category=Fully open models, Training=SFT, Instruction-tuned=true, Reasoning=Chain-of-thought (CoT)2026.03 | 63.1 | |
| Apertus-8BModel Category=Fully open models, Instruction-tuned=true, Reasoning=Chain-of-thought (CoT)2026.03 | 55.7 | |
| EuroLLM-9BModel Category=Fully open models, Instruction-tuned=true, Reasoning=Chain-of-thought (CoT)2026.03 | 55.2 | |
| Mistral-7BModel Category=Open weight models, Instruction-tuned=true, Reasoning=Chain-of-thought (CoT)2026.03 | 53.6 | |
| OLMo 2-7BModel Category=Fully open models, Instruction-tuned=true, Reasoning=Chain-of-thought (CoT)2026.03 | 45 | |
| Salamandra-7BModel Category=Fully open models, Instruction-tuned=true, Reasoning=Chain-of-thought (CoT)2026.03 | 37.7 |