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 |
|---|---|---|---|---|---|---|
| Hindi → Yue Chinese (Cantonese)HI · Hindi · हिन्दी→YUE · Yue Chinese (Cantonese) · 粵語 | 201 | Median 6.93 Mean 5.97 · 25th pct. 4.82 · 75th pct. 7.49 | Mean 0.203 25th pct. 0.202 · Median 0.206 · 75th pct. 0.210 | Mean 0.0 25th pct. 0.0 · Median 0.0 · 75th pct. 0.0 | Mean 0.1 25th pct. 0.0 · Median 0.0 · 75th pct. 0.0 | Mean 0.383 25th pct. 0.203 · Median 0.223 · 75th pct. 0.254 |
| Portuguese → Yue Chinese (Cantonese)PT · Portuguese · Português→YUE · Yue Chinese (Cantonese) · 粵語 | 63 | Median 8.18 Mean 7.17 · 25th pct. 7.42 · 75th pct. 8.79 | Mean 0.297 25th pct. 0.284 · Median 0.291 · 75th pct. 0.300 | Mean 0.0 25th pct. 0.0 · Median 0.0 · 75th pct. 0.0 | Mean 0.0 25th pct. 0.0 · Median 0.0 · 75th pct. 0.0 | Mean 0.208 25th pct. 0.197 · Median 0.207 · 75th pct. 0.212 |
| Modern Standard Arabic → Yue Chinese (Cantonese)AR · Modern Standard Arabic · العربية الفصحى→YUE · Yue Chinese (Cantonese) · 粵語 | 7 | Median 8.82 Mean 7.36 · 25th pct. 7.38 · 75th pct. 9.08 | Mean 0.299 25th pct. 0.297 · Median 0.299 · 75th pct. 0.302 | Mean 0.1 25th pct. 0.0 · Median 0.0 · 75th pct. 0.0 | Mean 0.3 25th pct. 0.0 · Median 0.0 · 75th pct. 0.0 | Mean 0.189 25th pct. 0.170 · Median 0.186 · 75th pct. 0.210 |
| Russian → Egyptian ArabicRU · Russian · Русский→ARZ · Egyptian Arabic · مصري | 14 | Median 8.71 Mean 7.69 · 25th pct. 8.03 · 75th pct. 9.04 | Mean 0.456 25th pct. 0.446 · Median 0.456 · 75th pct. 0.472 | Mean 0.7 25th pct. 0.1 · Median 0.2 · 75th pct. 0.3 | Mean 2.7 25th pct. 2.6 · Median 2.7 · 75th pct. 2.7 | Mean 0.893 25th pct. 0.857 · Median 0.883 · 75th pct. 0.911 |
| Portuguese → Egyptian ArabicPT · Portuguese · Português→ARZ · Egyptian Arabic · مصري | 47 | Median 8.36 Mean 7.39 · 25th pct. 7.16 · 75th pct. 8.85 | Mean 0.505 25th pct. 0.496 · Median 0.508 · 75th pct. 0.517 | Mean 5.1 25th pct. 4.5 · Median 4.5 · 75th pct. 6.5 | Mean 8.4 25th pct. 7.5 · Median 8.4 · 75th pct. 8.8 | Mean 0.768 25th pct. 0.751 · Median 0.778 · 75th pct. 0.798 |
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.