language benchmark
WMT24++ Leaderboard
WMT24++ is a multilingual machine translation benchmark covering 55 languages and dialects across literary, news, social, and speech domains.
Updated Sep 5, 2026
Models23
Model coverage23
MetricScore
EvidenceB
WMT24++ Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelNV | Score86.67% | Percentile100.00% | Participants23 | EvidenceC | Evaluated |
| Rank02 | ModelNV | Score86.20% | Percentile95.45% | Participants23 | EvidenceC | Evaluated |
| Rank03 | ModelAC | Score85.80% | Percentile90.91% | Participants23 | EvidenceC | Evaluated |
| Rank04 | ModelAC | Score84.60% | Percentile86.36% | Participants23 | EvidenceC | Evaluated |
| Rank05 | ModelAC | Score84.30% | Percentile81.82% | Participants23 | EvidenceC | Evaluated |
| Rank06 | ModelNV | Score83.70% | Percentile77.27% | Participants23 | EvidenceC | Evaluated |
| Rank07 | ModelCO | Score81.00% | Percentile72.73% | Participants23 | EvidenceC | Evaluated |
| Rank08 | ModelAC | Score78.90% | Percentile68.18% | Participants23 | EvidenceC | Evaluated |
| Rank09 | ModelAC | Score78.30% | Percentile63.64% | Participants23 | EvidenceC | Evaluated |
| Rank10 | ModelAC | Score77.60% | Percentile59.09% | Participants23 | EvidenceC | Evaluated |
| Rank11 | ModelAC | Score76.30% | Percentile54.55% | Participants23 | EvidenceC | Evaluated |
| Rank12 | ModelAC | Score72.60% | Percentile50.00% | Participants23 | EvidenceC | Evaluated |
| Rank13 | ModelAC | Score66.60% | Percentile45.45% | Participants23 | EvidenceC | Evaluated |
| Rank14 | ModelGO | Score53.40% | Percentile40.91% | Participants23 | EvidenceC | Evaluated |
| Rank15 | ModelGO | Score51.60% | Percentile36.36% | Participants23 | EvidenceC | Evaluated |
| Rank16 | ModelGO | Score50.10% | Percentile31.82% | Participants23 | EvidenceC | Evaluated |
| Rank17 | ModelGO | Score50.10% | Percentile27.27% | Participants23 | EvidenceC | Evaluated |
| Rank18 | ModelGO | Score46.80% | Percentile22.73% | Participants23 | EvidenceC | Evaluated |
| Rank19 | ModelAC | Score45.80% | Percentile18.18% | Participants23 | EvidenceC | Evaluated |
| Rank20 | ModelGO | Score42.70% | Percentile13.64% | Participants23 | EvidenceC | Evaluated |
| Rank21 | ModelGO | Score42.70% | Percentile9.09% | Participants23 | EvidenceC | Evaluated |
| Rank22 | ModelGO | Score35.90% | Percentile4.55% | Participants23 | EvidenceC | Evaluated |
| Rank23 | ModelAC | Score27.20% | Percentile0.00% | Participants23 | EvidenceC | Evaluated |
WMT24++ Highlights
The leading models and scores on this benchmark.
WMT24++ Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for WMT24++
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
What is WMT24++?
What WMT24++ measures and how its scores work.
WMT24++ expands the WMT24 dataset with human-written references and post-edits for evaluating machine translation systems and large language models.
It measures performance using the Score metric, reported as a ratio, across diverse linguistic contexts.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- WMT24++
- Modality
- text
- Primary category
- language
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- overall
LLM Stats. Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about WMT24++.
Which model scores highest on WMT24++?
Nemotron 3 Super (120B A12B) is currently ranked first with 86.67%.
What are the top three models on WMT24++?
The current leaders are Nemotron 3 Super (120B A12B) (86.67%), Nemotron 3 Nano (30B A3B) (86.20%), and Qwen3.7 Max (85.80%).
Which WMT24++ model has the lowest official input price?
Nemotron 3 Super (120B A12B) has the lowest matched official input price at $0.20 input / $0.80 output per 1M tokens.
Which models are fastest among WMT24++ results?
The fastest matched records are Gemma 3 27B (33.00 tok/s via DeepInfra), Gemma 3 12B (33.00 tok/s via DeepInfra), and Gemma 3 4B (33.00 tok/s via DeepInfra).
Does the highest score result prove overall model quality?
No. This benchmark measures one defined capability or task. The overall LLMBoard score uses a separate aggregation across eligible benchmark evidence.