Multitask Language Understanding on MMLU-ProX non-EU languages (test)
70.9AccuracyQwen-3-30B-A3B
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen-3-30B-A3BRelease Type=Open-weights, Model Category=Non-European2026.02 | 70.9 | |
| Qwen-3-32BRelease Type=Open-weights, Model Category=Non-European2026.02 | 69 | |
| Qwen-3-14BRelease Type=Open-weights, Model Category=Non-European2026.02 | 64.6 | |
| Llama-3.3-70BRelease Type=Open-weights, Model Category=Non-European2026.02 | 63.4 | |
| Mistral-3.2-24BRelease Type=Open-weights, Model Category=European2026.02 | 62.5 | |
| Gemma-3-27BRelease Type=Open-weights, Model Category=Non-European2026.02 | 58.8 | |
| OLMo-3.1-32BRelease Type=Fully-open, Model Category=Non-European2026.02 | 55.1 | |
| Gemma-3-12BRelease Type=Open-weights, Model Category=Non-European2026.02 | 51.8 | |
| EuroLLM-22B (new)Release Type=Fully-open, Model Category=European2026.02 | 43.8 | |
| OLMo-3-7BRelease Type=Fully-open, Model Category=Non-European2026.02 | 40.5 | |
| EuroLLM-9B (new)Release Type=Fully-open, Model Category=European2026.02 | 36.5 | |
| EuroLLM-22B (old)Release Type=Fully-open, Model Category=European2026.02 | 36.4 | |
| Apertus-70BRelease Type=Fully-open, Model Category=European2026.02 | 35.2 | |
| Llama-3.1-8BRelease Type=Open-weights, Model Category=Non-European2026.02 | 31.1 | |
| Apertus-8BRelease Type=Fully-open, Model Category=European2026.02 | 28.6 | |
| EuroLLM-9B (old)Release Type=Fully-open, Model Category=European2026.02 | 27.9 |