language benchmark
WMT23
The Eighth Conference on Machine Translation (WMT23) benchmark evaluating machine translation systems across 8 language pairs (14 translation directions) including general, biomedical, literary, and low-resource language translation tasks. Features specialized shared tasks for quality estimation, metrics evaluation, sign language translation, and discourse-level literary translation with professional human assessment.
Updated Aug 7, 2026
Published models4
Registry coverage4
MetricScore
EvidenceB
WMT23 leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | GO | 75.1% | 100.0% | 4 | C | |
| 2 | GO | 74.1% | 66.7% | 4 | C | |
| 3 | GO | 72.6% | 33.3% | 4 | C | |
| 4 | GO | 71.7% | 0.0% | 4 | B |
Score distribution
Top published rows on the benchmark's original scale.
WMT23 highlights
The top published results on this benchmark's own scale.
What is WMT23?
Definition and scoring fields from the benchmark registry.
The Eighth Conference on Machine Translation (WMT23) benchmark evaluating machine translation systems across 8 language pairs (14 translation directions) including general, biomedical, literary, and low-resource language translation tasks. Features specialized shared tasks for quality estimation, metrics evaluation, sign language translation, and discourse-level literary translation with professional human assessment.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- WMT23
- Modality
- text
- Primary category
- language
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- wmt23|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about WMT23.
Which model scores highest on WMT23?
Gemini 1.5 Pro is currently ranked first with 75.1%.
What does WMT23 measure?
The Eighth Conference on Machine Translation (WMT23) benchmark evaluating machine translation systems across 8 language pairs (14 translation directions) including general, biomedical, literary, and low-resource language translation tasks. Features specialized shared tasks for quality estimation, metrics evaluation, sign language translation, and discourse-level literary translation with professional human assessment.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
4 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is preserved as source-native evidence but is not eligible for the current overall score.