Science Question Answering on ARC Arabic translated (test)
89ARC ScoreTrinity Large (MoE)
Evaluation results
| Method | Links | |
|---|---|---|
| Trinity Large (MoE)Zero-shot=true2026.02 | 89 | |
| Qwen3 8BZero-shot=true2026.02 | 83 | |
| Qwen3 4BZero-shot=true2026.02 | 78 | |
| Datology 8BZero-shot=true2026.02 | 67 | |
| LFM2.5 1.2BZero-shot=true2026.02 | 62 | |
| Granite-4.0 MicroZero-shot=true2026.02 | 60 | |
| Datology 3BZero-shot=true2026.02 | 59 | |
| SmolLM3 3BZero-shot=true2026.02 | 59 | |
| Llama-3.1 8BZero-shot=true2026.02 | 55 | |
| Llama-3.2 3BZero-shot=true2026.02 | 46 | |
| Llama-3.2 1BZero-shot=true2026.02 | 27 |