Question Answering on ARC Spanish (test)
0.92AccuracyTrinity Large
Evaluation Results
| Method | Links | |
|---|---|---|
| Trinity LargeArchitecture=MoE, Protocol=Zero-shot2026.02 | 0.92 | |
| Qwen3Parameters=8B, Protocol=Zero-shot2026.02 | 0.9 | |
| Qwen3Parameters=4B, Protocol=Zero-shot2026.02 | 0.86 | |
| DatologyParameters=8B, Protocol=Zero-shot2026.02 | 0.73 | |
| Granite-4.0Parameters=Micro, Protocol=Zero-shot2026.02 | 0.7 | |
| Llama-3.1Parameters=8B, Protocol=Zero-shot2026.02 | 0.67 | |
| SmolLM3Parameters=3B, Protocol=Zero-shot2026.02 | 0.66 | |
| LFM2.5Parameters=1.2B, Protocol=Zero-shot2026.02 | 0.66 | |
| DatologyParameters=3B, Protocol=Zero-shot2026.02 | 0.63 | |
| Llama-3.2Parameters=3B, Protocol=Zero-shot2026.02 | 0.56 | |
| Llama-3.2Parameters=1B, Protocol=Zero-shot2026.02 | 0.31 |