Bornholm Parallel
Models match Bornholmsk sentences to their Danish translations by comparing sentence vectors. The test contains 500 sentence pairs.
Google reports a Gemini Embedding 2 score of 64.14. The original Danish ranking publishes separate Bornholm scores for 18 models.
- Released
- July 2023
- Kind
- Translation
- Level
- School
- Scored
- Checked automatically
- Answers
- Prediction
- Tasks
- All public
- Measures
- Capability
- Subject
- Languages and literature
- Language
- Danish
- Interaction
- Single turn
- Items from
- Real work and use
- Refresh
- One fixed set
- Run by
- Anyone
multilingual-e5-base tops the leaderboard with 46.4%
Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders.
- 46.4 multilingual-e5-base
- 44.2 multilingual-e5-large
- 43.9 multilingual-e5-small
- 40.3 e5-small
- 40.1 e5-large
- 40.1 e5-base
- 29.7 all-MiniLM-L6-v2
- 19.7 paraphrase-multilingual-MiniLM-L12-v2
- 15.9 dfm-sentence-encoder-large-1
- 14.1 sentence-bert-swedish-cased
- 11.7 dfm-encoder-large-v1
- 9.88 nb-bert-base
- 6.60 bert-base-swedish-cased
- 6.34 DanskBERT
- 4.53 nb-bert-large
- 4.42 xlm-roberta-base
- 1.44 electra-small-nordic
- 0.85 electra-small-swedish-cased-discriminator
- 1 multilingual-e5-base 46.4%
- 2 multilingual-e5-large 44.2%
- 3 multilingual-e5-small 43.9%
- 4 e5-small 40.3%
- 5 e5-large 40.1%
- 6 e5-base 40.1%
- 7 all-MiniLM-L6-v2 29.7%
- 8 paraphrase-multilingual-MiniLM-L12-v2 19.7%
- 9 dfm-sentence-encoder-large-1 15.9%
- 10 sentence-bert-swedish-cased 14.1%
- 11 dfm-encoder-large-v1 11.7%
- 12 nb-bert-base 9.88%
- 13 bert-base-swedish-cased 6.60%
- 14 DanskBERT 6.34%
- 15 nb-bert-large 4.53%
- 16 xlm-roberta-base 4.42%
- 17 electra-small-nordic 1.44%
- 18 electra-small-swedish-cased-discriminator 0.85%
Source: Official Bornholm Parallel leaderboard
| # | Entry | Score | Setting | Date | Link | Source |
|---|---|---|---|---|---|---|
| 1 | multilingual-e5-base | 46.4% | huggingface.co | Official leaderboard | ||
| 2 | multilingual-e5-large | 44.2% | huggingface.co | Official leaderboard | ||
| 3 | multilingual-e5-small | 43.9% | huggingface.co | Official leaderboard | ||
| 4 | e5-small | 40.3% | huggingface.co | Official leaderboard | ||
| 5 | e5-large | 40.1% | huggingface.co | Official leaderboard | ||
| 6 | e5-base | 40.1% | huggingface.co | Official leaderboard | ||
| 7 | all-MiniLM-L6-v2 | 29.7% | huggingface.co | Official leaderboard | ||
| 8 | paraphrase-multilingual-MiniLM-L12-v2 | 19.7% | huggingface.co | Official leaderboard | ||
| 9 | dfm-sentence-encoder-large-1 | 15.9% | huggingface.co | Official leaderboard | ||
| 10 | sentence-bert-swedish-cased | 14.1% | huggingface.co | Official leaderboard | ||
| 11 | dfm-encoder-large-v1 | 11.7% | huggingface.co | Official leaderboard | ||
| 12 | nb-bert-base | 9.88% | huggingface.co | Official leaderboard | ||
| 13 | bert-base-swedish-cased | 6.60% | huggingface.co | Official leaderboard | ||
| 14 | DanskBERT | 6.34% | huggingface.co | Official leaderboard | ||
| 15 | nb-bert-large | 4.53% | huggingface.co | Official leaderboard | ||
| 16 | xlm-roberta-base | 4.42% | huggingface.co | Official leaderboard | ||
| 17 | electra-small-nordic | 1.44% | huggingface.co | Official leaderboard | ||
| 18 | electra-small-swedish-cased-discriminator | 0.85% | huggingface.co | Official leaderboard |
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
More in Multilingual
- MMMLU In use MMLU professionally translated into 14 languages. Multiple-choice accuracy. Claude Mythos Preview 92.7%
- WMT24++ In use English-to-55-language translation of WMT24 test data. Scored with automatic metrics such as MetricX or xCOMET.
K-EXAONE 90.5%
- MMLU-ProX In use MMLU-Pro translated into up to 29 languages with expert review. Multiple-choice accuracy per language.
Qwen3.7-Max 87%
- CLIcK In use Models answer 1,995 multiple-choice questions on Korean language and culture. Questions come from official Korean exams and textbooks in eleven categories.
Solar Open 2 90.7%
- Global MMLU Lite 3.0 In use Models answer 400 human-translated MMLU questions per language in 23 languages. Version 3.0 adds Czech, Hungarian, Oriya, Slovak and Tajik and updates Italian. Accuracy. Claude Opus 5 95.7%
- MMMB In use Models answer multiple-choice questions about images in six languages across 15 categories. The set has about 12,000 questions. LFM2.5-VL-3B 83%
Data audit
All 8 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
- Pass Independent runs Scores from the official leaderboard (18 entries).
- Pass Models scored 18 entries on its leaderboard, none a model we track.
- Note Newest score No scores yet.
- Note Frontier models No models scored yet.
- Note Headroom No scores yet.
Leaderboard import 3 pass
- Pass Importer We read the board on datawrapper.dwcdn.net: 18 entries at the last read, none a model we track.
- Pass Last read November 20th, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.