ltzGLUE
Models perform language understanding tasks in Luxembourgish, including entailment, grammar judgments, sentiment analysis, and named entity recognition.
The authors test grammar, sentiment, entailment, named entities, and other tasks. EuroEval documents eight configurations for these tasks.
- Made by
- Alistair Plum and colleagues
- Released
- April 2026
- Kind
- Reasoning in many languages
- Level
- School
- Scored
- Checked automatically
- Answers
- Multiple choice
- Tasks
- All public
- Measures
- Capability
- Subject
- Languages and literature
- Language
- Luxembourgish
- Interaction
- Single turn
- Refresh
- One fixed set
- Run by
- Anyone
- Strongest model tested
- GPT-5 mini #92 on the capability index
- Last model added
- August 2025
- Scores refreshed
- November 20th, 2026
mmBRT tops the leaderboard with 79.8 GPT-5 mini leads with 69.4 among models released by August 2025
Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. The paper reports no overall F1, so the headline is the plain mean of the eight test F1 scores. 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. The best model we track is GPT-5 mini, at 69.4. This view shows only models we track, each at its best.
The newest model with a score here came out in August 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 1 mmBRT Supervised fine-tuning 79.8
- 2 LxBRT Supervised fine-tuning 78.3
- 3 ltzE1m Supervised fine-tuning 75.9
- 4 ltzE1b Supervised fine-tuning 71.4
- 5 mBRT Supervised fine-tuning 70.7
- 6 GPT-5 mini Zero-shot 69.4
- 7 XLMR Supervised fine-tuning 65.9
- 8
Qwen3-235B-A22B Zero-shot 65.4
- 9
Gemma 3 27B Zero-shot 52.0
- 10
Llama 3.3 70B Zero-shot 43.4
Source: The ltzGLUE paper
- 1 GPT-5 mini Zero-shot 69.4
- 2
Qwen3-235B-A22B Zero-shot 65.4
- 3
Gemma 3 27B Zero-shot 52.0
- 4
Llama 3.3 70B Zero-shot 43.4
Source: The ltzGLUE paper
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | mmBRT | 79.8 | Supervised fine-tuning | Official leaderboard | |
| 2 | LxBRT | 78.3 | Supervised fine-tuning | Official leaderboard | |
| 3 | ltzE1m | 75.9 | Supervised fine-tuning | Official leaderboard | |
| 4 | ltzE1b | 71.4 | Supervised fine-tuning | Official leaderboard | |
| 5 | mBRT | 70.7 | Supervised fine-tuning | Official leaderboard | |
| 6 | GPT-5 mini | 69.4 | Zero-shot | Aug 2025 | Official leaderboard |
| 7 | XLMR | 65.9 | Supervised fine-tuning | Official leaderboard | |
| 8 | | 65.4 | Zero-shot | Apr 2025 | Official leaderboard |
| 9 | | 52.0 | Zero-shot | Mar 2025 | Official leaderboard |
| 10 | | 43.4 | Zero-shot | Dec 2024 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-5 mini | 69.4 | Zero-shot | Aug 2025 | Official leaderboard |
| 2 | | 65.4 | Zero-shot | Apr 2025 | Official leaderboard |
| 3 | | 52.0 | Zero-shot | Mar 2025 | Official leaderboard |
| 4 | | 43.4 | Zero-shot | Dec 2024 | Official leaderboard |
Rankings by part
We read 8 parts from the benchmark’s paper, each ranking models on one part of the test, such as Headline acceptability, Sentiment analysis and Binary linguistic acceptability. GPT-5 mini leads 7 of them, and Qwen3-235B-A22B leads the other one. Arrows show how far a model moved from its overall rank among the models we track.
Headline acceptability
4 models we track have a score here.
- 1 GPT-5 mini 88.9
- 2
Qwen3-235B-A22B 86.1
- 3
Gemma 3 27B 80.5
- 4
Llama 3.3 70B 77.7
Sentiment analysis
4 models we track have a score here.
- 1 GPT-5 mini 56.4
- 2
Gemma 3 27B ↑1 55.6
- 3
Qwen3-235B-A22B ↓1 51.6
- 4
Llama 3.3 70B 45.8
Binary linguistic acceptability
4 models we track have a score here.
- 1 GPT-5 mini 75.2
- 2
Qwen3-235B-A22B 67.8
- 3
Gemma 3 27B 58.7
- 4
Llama 3.3 70B 41.4
Multiclass linguistic acceptability
4 models we track have a score here.
- 1 GPT-5 mini 51.5
- 2
Gemma 3 27B ↑1 43.3
- 3
Qwen3-235B-A22B ↓1 39.7
- 4
Llama 3.3 70B 11.2
Named entity recognition
4 models we track have a score here.
- 1
Qwen3-235B-A22B ↑1 70.7
- 2 GPT-5 mini ↓1 67.2
- 3
Llama 3.3 70B ↑1 50.5
- 4
Gemma 3 27B ↓1 48.4
Topic classification
4 models we track have a score here.
- 1 GPT-5 mini 89.3
- 2
Qwen3-235B-A22B 88.4
- 3
Gemma 3 27B 48.7
- 4
Llama 3.3 70B 12.5
Intent detection
4 models we track have a score here.
- 1 GPT-5 mini 38.3
- 2
Llama 3.3 70B ↑2 35.8
- 3
Qwen3-235B-A22B ↓1 34.9
- 4
Gemma 3 27B ↓1 7.25
Textual entailment
4 models we track have a score here.
- 1 GPT-5 mini 88.5
- 2
Qwen3-235B-A22B 84.2
- 3
Gemma 3 27B 73.1
- 4
Llama 3.3 70B 72.0
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: The ltzGLUE paper
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official ltzGLUE 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
8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 15 months ago, in August 2025, for GPT-5 mini. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
- Pass Independent runs Scores from the benchmark’s paper (10 entries).
- Pass Models scored 4 models we track, plus 6 other entries on its leaderboard.
- Warning Newest score The newest came 15 months ago, in August 2025, for GPT-5 mini.
- Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5 mini (#92).
- Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on arxiv.org: 4 rows for models we track and 6 other entries at the last read.
- 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.