Machine Translation on Chintang (ctn) (test)
12.59BLEUgemma-4-31B-it
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| gemma-4-31B-itReasoning=true2026.06 | 12.59 | 36.53 | 73.46 | 65.14 | |
| Qwen3-14BReasoning=true2026.06 | 10.33 | 34.07 | 67.07 | 51.98 | |
| gemma-4-31B-itReasoning=false2026.06 | 9.96 | 31.61 | 63.13 | 43.24 | |
| Qwen3-8BReasoning=true2026.06 | 7.02 | 31.19 | 64.8 | 45.55 | |
| gemma-4-E4B-itReasoning=true2026.06 | 6.96 | 29.75 | 65.04 | 49.99 | |
| Qwen3-14BReasoning=false2026.06 | 5.25 | 26.5 | 56.1 | 32.91 | |
| gemma-4-E2B-itReasoning=false2026.06 | 3.79 | 22.51 | 49.2 | 31.77 | |
| gemma-4-E2B-itReasoning=true2026.06 | 3.37 | 24.08 | 51.26 | 41.54 | |
| Qwen3-8BReasoning=false2026.06 | 3.28 | 20.84 | 46.23 | 24.43 | |
| gemma-4-E4B-itReasoning=false2026.06 | 1.39 | 17.86 | 53.72 | 31.48 | |
| Qwen3-4B-ThinkingReasoning=true2026.06 | 1.11 | 13.12 | 63.77 | 48.56 | |
| Qwen3-4B-ThinkingReasoning=false2026.06 | 0.22 | 5.86 | 44.03 | 25.14 |