Machine Translation on WMT Xx ⇒ En 24++
85sCMT ScoreLOONG
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| LOONGBackbone=Qwen3-14B2026.05 | 85 | 82.7 | 85.3 | 84.4 | |
| Doc2DocBackbone=Qwen3-14B2026.05 | 84.9 | 82.6 | 85.1 | 84.2 | |
| LOONGBackbone=Qwen3-8B2026.05 | 84.8 | 82.5 | — | 83.6 | |
| SegmentBackbone=Qwen3-14B2026.05 | 84.8 | 82.5 | 82.6 | 83.3 | |
| DELTABackbone=Qwen3-14B2026.05 | 84.7 | 82.4 | 84.3 | 83.8 | |
| LOONGBackbone=Qwen2.5-7B2026.05 | 84.4 | 82.1 | 82 | 82.9 | |
| SentenceBackbone=Qwen3-14B2026.05 | 84.3 | 81.8 | 80.4 | 82.2 | |
| DELTABackbone=Qwen3-8B2026.05 | 84.2 | 81.9 | 81.1 | 82.4 | |
| LOONGBackbone=Llama3.1-8B2026.05 | 84.1 | 81.7 | 82.3 | 82.7 | |
| DELTABackbone=Qwen2.5-7B2026.05 | 83.9 | 81.4 | 80.3 | 81.9 | |
| Doc2DocBackbone=Qwen3-8B2026.05 | 83.8 | 80.4 | 78.6 | 80.9 | |
| SegmentBackbone=Llama3.1-8B2026.05 | 83.6 | 80.8 | 74.6 | 79.6 | |
| Doc2DocBackbone=Llama3.1-8B2026.05 | 83.6 | 81 | 76.7 | 80.4 | |
| SentenceBackbone=Qwen3-8B2026.05 | 83.5 | 80.9 | 75.8 | 80 | |
| SegmentBackbone=Qwen3-8B2026.05 | 83.5 | 79.3 | 74.2 | 79 | |
| Doc2DocBackbone=Qwen2.5-7B2026.05 | 83 | 78.8 | 74.8 | 78.8 | |
| SegmentBackbone=Qwen2.5-7B2026.05 | 82.7 | 78.2 | 73.8 | 78.2 | |
| SentenceBackbone=Llama3.1-8B2026.05 | 80.5 | 77.7 | 69.6 | 75.9 | |
| DELTABackbone=Llama3.1-8B2026.05 | 79.7 | 76.7 | 75.2 | 77.2 | |
| SentenceBackbone=Qwen2.5-7B2026.05 | 77.8 | 74.6 | 57.3 | 69.9 |