@inproceedings{devine-etal-2026-gaeleval,
title = "{G}ael{E}val: Benchmarking {LLM} Performance for {S}cottish {G}aelic",
author = "Devine, Peter and
Lamb, William and
Alex, Beatrice and
Ezeani, Ignatius and
Knight, Dawn and
{\'O} Meachair, M{\'i}che{\'a}l J. and
Rayson, Paul and
Wynne, Martin",
editor = "Montejo-Raez, Arturo and
Grisot, Cristina and
Blochowiak, Joanna and
Ljube{\v{s}}i{\'c}, Nikola and
Battaner, Elena and
Rigau, German",
booktitle = "Proceedings of Shaping Multilingual, Multimodal {AI} for the Social Sciences and Humanities ({LLM}s4{SSH}) @ {LREC} 2026",
month = may,
year = "2026",
address = "Palma de Mallorca (Spain)",
publisher = "ELRA Language Resources Association (ELRA)",
url = "https://aclanthology.org/2026.llms4ssh-1.8/",
doi = "10.63317/2yqmx3es6ad5",
pages = "73--85",
abstract = "Multilingual large language models (LLMs) often exhibit emergent `shadow' capabilities in languages without official support, yet their performance on these languages remains uneven and under-measured. This is particularly acute for morphosyntactically rich minority languages such as Scottish Gaelic, where translation benchmarks fail to capture structural competence. We introduce GaelEval, the first multi-dimensional benchmark for Gaelic, comprising: (i) an expert-authored morphosyntactic MCQA task; (ii) a culturally-grounded translation benchmark and (iii) a large-scale cultural knowledge Q{\&}A task. Evaluating 19 LLMs against a fluent-speaker human baseline (n = 30), we find that Gemini 3 Pro Preview achieves 83.3{\%} accuracy on the linguistic task, surpassing the human baseline (78.1{\%}). Proprietary models consistently outperform open-weight systems, and in-language (Gaelic) prompting yields a small but stable advantage (+2.4pp). On the cultural task, leading models exceed 90{\%} accuracy, though most systems perform worse under Gaelic prompting and absolute scores are inflated relative to the manual benchmark. Overall, GaelEval reveals that frontier models achieve above-human performance on several dimensions of Gaelic grammar, demonstrates the effect of Gaelic prompting and shows a consistent performance gap favouring proprietary over open-weight models."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="devine-etal-2026-gaeleval">
<titleInfo>
<title>GaelEval: Benchmarking LLM Performance for Scottish Gaelic</title>
</titleInfo>
<name type="personal">
<namePart type="given">Peter</namePart>
<namePart type="family">Devine</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">William</namePart>
<namePart type="family">Lamb</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Beatrice</namePart>
<namePart type="family">Alex</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ignatius</namePart>
<namePart type="family">Ezeani</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dawn</namePart>
<namePart type="family">Knight</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mícheál</namePart>
<namePart type="given">J</namePart>
<namePart type="family">Ó Meachair</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Paul</namePart>
<namePart type="family">Rayson</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Martin</namePart>
<namePart type="family">Wynne</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-05</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of Shaping Multilingual, Multimodal AI for the Social Sciences and Humanities (LLMs4SSH) @ LREC 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Arturo</namePart>
<namePart type="family">Montejo-Raez</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Cristina</namePart>
<namePart type="family">Grisot</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Joanna</namePart>
<namePart type="family">Blochowiak</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nikola</namePart>
<namePart type="family">Ljubešić</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Elena</namePart>
<namePart type="family">Battaner</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">German</namePart>
<namePart type="family">Rigau</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>ELRA Language Resources Association (ELRA)</publisher>
<place>
<placeTerm type="text">Palma de Mallorca (Spain)</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Multilingual large language models (LLMs) often exhibit emergent ‘shadow’ capabilities in languages without official support, yet their performance on these languages remains uneven and under-measured. This is particularly acute for morphosyntactically rich minority languages such as Scottish Gaelic, where translation benchmarks fail to capture structural competence. We introduce GaelEval, the first multi-dimensional benchmark for Gaelic, comprising: (i) an expert-authored morphosyntactic MCQA task; (ii) a culturally-grounded translation benchmark and (iii) a large-scale cultural knowledge Q&A task. Evaluating 19 LLMs against a fluent-speaker human baseline (n = 30), we find that Gemini 3 Pro Preview achieves 83.3% accuracy on the linguistic task, surpassing the human baseline (78.1%). Proprietary models consistently outperform open-weight systems, and in-language (Gaelic) prompting yields a small but stable advantage (+2.4pp). On the cultural task, leading models exceed 90% accuracy, though most systems perform worse under Gaelic prompting and absolute scores are inflated relative to the manual benchmark. Overall, GaelEval reveals that frontier models achieve above-human performance on several dimensions of Gaelic grammar, demonstrates the effect of Gaelic prompting and shows a consistent performance gap favouring proprietary over open-weight models.</abstract>
<identifier type="citekey">devine-etal-2026-gaeleval</identifier>
<identifier type="doi">10.63317/2yqmx3es6ad5</identifier>
<location>
<url>https://aclanthology.org/2026.llms4ssh-1.8/</url>
</location>
<part>
<date>2026-05</date>
<extent unit="page">
<start>73</start>
<end>85</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T GaelEval: Benchmarking LLM Performance for Scottish Gaelic
%A Devine, Peter
%A Lamb, William
%A Alex, Beatrice
%A Ezeani, Ignatius
%A Knight, Dawn
%A Ó Meachair, Mícheál J.
%A Rayson, Paul
%A Wynne, Martin
%Y Montejo-Raez, Arturo
%Y Grisot, Cristina
%Y Blochowiak, Joanna
%Y Ljubešić, Nikola
%Y Battaner, Elena
%Y Rigau, German
%S Proceedings of Shaping Multilingual, Multimodal AI for the Social Sciences and Humanities (LLMs4SSH) @ LREC 2026
%D 2026
%8 May
%I ELRA Language Resources Association (ELRA)
%C Palma de Mallorca (Spain)
%F devine-etal-2026-gaeleval
%X Multilingual large language models (LLMs) often exhibit emergent ‘shadow’ capabilities in languages without official support, yet their performance on these languages remains uneven and under-measured. This is particularly acute for morphosyntactically rich minority languages such as Scottish Gaelic, where translation benchmarks fail to capture structural competence. We introduce GaelEval, the first multi-dimensional benchmark for Gaelic, comprising: (i) an expert-authored morphosyntactic MCQA task; (ii) a culturally-grounded translation benchmark and (iii) a large-scale cultural knowledge Q&A task. Evaluating 19 LLMs against a fluent-speaker human baseline (n = 30), we find that Gemini 3 Pro Preview achieves 83.3% accuracy on the linguistic task, surpassing the human baseline (78.1%). Proprietary models consistently outperform open-weight systems, and in-language (Gaelic) prompting yields a small but stable advantage (+2.4pp). On the cultural task, leading models exceed 90% accuracy, though most systems perform worse under Gaelic prompting and absolute scores are inflated relative to the manual benchmark. Overall, GaelEval reveals that frontier models achieve above-human performance on several dimensions of Gaelic grammar, demonstrates the effect of Gaelic prompting and shows a consistent performance gap favouring proprietary over open-weight models.
%R 10.63317/2yqmx3es6ad5
%U https://aclanthology.org/2026.llms4ssh-1.8/
%U https://doi.org/10.63317/2yqmx3es6ad5
%P 73-85
Markdown (Informal)
[GaelEval: Benchmarking LLM Performance for Scottish Gaelic](https://aclanthology.org/2026.llms4ssh-1.8/) (Devine et al., LLMs4SSH 2026)
ACL
- Peter Devine, William Lamb, Beatrice Alex, Ignatius Ezeani, Dawn Knight, Mícheál J. Ó Meachair, Paul Rayson, and Martin Wynne. 2026. GaelEval: Benchmarking LLM Performance for Scottish Gaelic. In Proceedings of Shaping Multilingual, Multimodal AI for the Social Sciences and Humanities (LLMs4SSH) @ LREC 2026, pages 73–85, Palma de Mallorca (Spain). ELRA Language Resources Association (ELRA).