Loading benchmarks…
Benchmark dashboard
Automated translation metrics summarized for 335 language directions.
Sources — Source segments come from curated, public-facing parallel corpora and benchmark-style collections mixed for broad coverage. They are not model-generated: originals and references are assembled upstream as parallel text only; this dashboard does not redistribute raw corpora.
Mixed parallel corpora — Parallel segments spanning diverse genres and corpus scales. Each item includes a reference translation that is certified and human-verified—not produced by the models being benchmarked.
Language pairi | Rankingi | Fluencyi | COMETi | BLEUi | sacreBLEUi | Len ratioi |
|---|---|---|---|---|---|---|
| Mandarin Chinese → VietnameseZH · Mandarin Chinese · 普通话→VI · Vietnamese · Tiếng Việt | 324 | Median 4.01 Mean 3.72 · 25th pct. 3.65 · 75th pct. 4.49 | Mean 0.870 25th pct. 0.872 · Median 0.876 · 75th pct. 0.881 | Mean 36.9 25th pct. 35.6 · Median 38.1 · 75th pct. 39.5 | Mean 38.5 25th pct. 37.7 · Median 39.5 · 75th pct. 41.4 | Mean 50.573 25th pct. 50.167 · Median 50.889 · 75th pct. 51.000 |
| Mandarin Chinese → Modern Standard ArabicZH · Mandarin Chinese · 普通话→AR · Modern Standard Arabic · العربية الفصحى | 302 | Median 4.82 Mean 4.31 · 25th pct. 4.31 · 75th pct. 5.00 | Mean 0.831 25th pct. 0.831 · Median 0.847 · 75th pct. 0.850 | Mean 26.3 25th pct. 20.7 · Median 25.0 · 75th pct. 33.2 | Mean 28.0 25th pct. 20.8 · Median 27.5 · 75th pct. 35.8 | Mean 41.474 25th pct. 41.000 · Median 41.000 · 75th pct. 42.500 |
| Mandarin Chinese → Iranian Persian (Farsi)ZH · Mandarin Chinese · 普通话→FA · Iranian Persian (Farsi) · فارسی | 313 | Median 4.28 Mean 4.03 · 25th pct. 3.94 · 75th pct. 4.74 | Mean 0.850 25th pct. 0.857 · Median 0.863 · 75th pct. 0.868 | Mean 16.2 25th pct. 14.6 · Median 16.9 · 75th pct. 17.6 | Mean 16.8 25th pct. 15.7 · Median 17.2 · 75th pct. 18.5 | Mean 35.053 25th pct. 33.778 · Median 34.889 · 75th pct. 35.833 |
| Mandarin Chinese → IndonesianZH · Mandarin Chinese · 普通话→ID · Indonesian · Bahasa Indonesia | 309 | Median 4.39 Mean 4.00 · 25th pct. 4.02 · 75th pct. 4.76 | Mean 0.877 25th pct. 0.877 · Median 0.881 · 75th pct. 0.882 | Mean 13.8 25th pct. 11.2 · Median 13.8 · 75th pct. 15.6 | Mean 17.8 25th pct. 15.8 · Median 17.3 · 75th pct. 20.0 | Mean 32.743 25th pct. 31.944 · Median 32.778 · 75th pct. 33.167 |
| Mandarin Chinese → TurkishZH · Mandarin Chinese · 普通话→TR · Turkish · Türkçe | 310 | Median 4.36 Mean 4.00 · 25th pct. 3.87 · 75th pct. 4.53 | Mean 0.852 25th pct. 0.849 · Median 0.853 · 75th pct. 0.858 | Mean 8.9 25th pct. 7.7 · Median 9.1 · 75th pct. 9.9 | Mean 10.5 25th pct. 9.0 · Median 10.7 · 75th pct. 11.9 | Mean 27.655 25th pct. 27.278 · Median 27.778 · 75th pct. 28.222 |
This release reports translation quality by language pair: medians and spread of automated scores (fluency, COMET, BLEU, sacreBLEU, length ratio) aggregated across evaluated directions. Run metadata describes recipes, metrics, and segment counts.