Machine Translation on NH 3.0 (test)
40.24BLEUOurs
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Ours2026.01 | 40.24 | 55.89 | |
| Azure Translator2026.01 | 34.76 | 52.11 | |
| OursEvaluation protocol=-2026.01 | 21.08 | 46.34 | |
| GPT-5-ReasoningEvaluation Protocol=5-shot2026.01 | 18.02 | 38.17 | |
| Azure TranslatorEvaluation protocol=-2026.01 | 17.17 | 44.69 | |
| GPT-4oEvaluation Protocol=5-shot2026.01 | 15.35 | 33.56 | |
| GPT-5-ReasoningEvaluation Protocol=0-shot2026.01 | 14.44 | 34.99 | |
| GPT-4.1Evaluation Protocol=5-shot2026.01 | 14.16 | 32.48 | |
| GPT-4oEvaluation Protocol=0-shot2026.01 | 11.29 | 29.9 | |
| GPT-5-ReasoningEvaluation protocol=5-shot2026.01 | 6.61 | 28.31 | |
| GPT-5-ReasoningEvaluation protocol=0-shot2026.01 | 5.68 | 24.42 | |
| GPT-4.1Evaluation Protocol=0-shot2026.01 | 5.59 | 21.69 | |
| GPT-4.1Evaluation protocol=5-shot2026.01 | 3.34 | 21.23 | |
| GPT-4oEvaluation protocol=5-shot2026.01 | 2.56 | 17.46 | |
| GPT-4oEvaluation protocol=0-shot2026.01 | 1.28 | 12.01 | |
| GPT-4.1Evaluation protocol=0-shot2026.01 | 0.97 | 11.78 |