Skip to content

Benchmark

LuxIT base RTE F1-macro

Multilingual benchmarks compare model performance across languages. Higher scores are better.

Models and systems measured
14
Scores recorded
14
Sources
1
IQ weight
No weight published

Score distribution

0.445
Mean
0.449
Median
0.208
Std dev
0.142
10th percentile
0.744
90th percentile
0.139
Min
0.1390.775
The band covers the middle 80% of scores. The line marks the median.

Results

The source marks these results as descriptive, not capability scores, so we list them by name instead of ranking them.

Model or systemScore
Apertus-8B-Instruct-2509table: 10, column: RTE, system: apertus-8b· LuxIT (arXiv:2510.24434v4) data0.686
EuroLLM-1.7B-Instructtable: 10, column: RTE, system: eurollm-1.7b· LuxIT (arXiv:2510.24434v4) data0.139
Gemma-3-12B-ITtable: 10, column: RTE, system: gemma-3-12b· LuxIT (arXiv:2510.24434v4) data0.744
Gemma-3-1B-ITtable: 10, column: RTE, system: gemma-3-1b· LuxIT (arXiv:2510.24434v4) data0.53
GLM-4-9B-0414table: 10, column: RTE, system: glm-4-9b· LuxIT (arXiv:2510.24434v4) data0.775
Llama-3.1-8B-Instructtable: 10, column: RTE, system: llama-3.1-8b· LuxIT (arXiv:2510.24434v4) data0.332
Llama-3.2-1B-Instructtable: 10, column: RTE, system: llama-3.2-1b· LuxIT (arXiv:2510.24434v4) data0.251
Ministral-3-3B-Instruct-2512table: 10, column: RTE, system: ministral-3-3b· LuxIT (arXiv:2510.24434v4) data0.392
Mistral-7B-Instruct-v0.3table: 10, column: RTE, system: mistral-7b· LuxIT (arXiv:2510.24434v4) data0.548
Olmo-3-7B-Instructtable: 10, column: RTE, system: olmo-3-7b· LuxIT (arXiv:2510.24434v4) data0.142
Phi-4table: 10, column: RTE, system: phi-4· LuxIT (arXiv:2510.24434v4) data0.449
Qwen2.5-0.5B-Instructtable: 10, column: RTE, system: qwen2.5-0.5b· LuxIT (arXiv:2510.24434v4) data0.276
Qwen2.5-1.5B-Instructtable: 10, column: RTE, system: qwen2.5-1.5b· LuxIT (arXiv:2510.24434v4) data0.31
Qwen2.5-7B-Instructtable: 10, column: RTE, system: qwen2.5-7b· LuxIT (arXiv:2510.24434v4) data0.658

Data credits: Julian Valline, Cedric Lothritz, Siwen Guo and Jordi Cabot. LuxIT: A Luxembourgish Instruction Tuning Dataset from Monolingual Seed Data. arXiv:2510.24434v4. CC BY 4.0 article numeric/protocol extract; restructured, printed values unchanged. Task data, private fine-tunes, exams and weights excluded. (licence)