GaelEval - Linguistic Competence (Gaelic Prompt): leaderboard
GaelEval (2026) is a Scottish Gaelic benchmark from the DR-LIB project. This board is accuracy in percent on its 120-item multiple-choice test of Gaelic morphosyntax and idiom, written by a Gaelic linguist across 16 grammatical categories, answered with a Gaelic-language system prompt; higher is better.
Metric: Accuracy (%) on the 120-item Gaelic morphosyntactic MCQA with a Gaelic-language system prompt, exact string match, single call per item, zero-shot; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 19 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 83.3 | #119 |
| 2 | Gemini 3 Flash (Preview) | 79.2 | #152 |
| 3 | GPT-5 | 69.2 | #178 |
| 4 | Gemini 2.5 Flash | 61.7 | #463 |
| 5 | Claude Opus 4.6 | 59.2 | #113 |
| 6 | GPT-4o | 50 | #751 |
| 7 | Claude Haiku 4.5 | 47.5 | #583 |
| 8 | GPT-5 Mini | 47.5 | #364 |
| 9 | DeepSeek R1 | 45 | #482 |
| 10 | Llama 4 Maverick | 45 | #1110 |
| 11 | GPT-5.2 | 43.3 | #185 |
| 12 | GPT-4.1 | 42.5 | #505 |
| 13 | GPT-5 Nano | 36.7 | #906 |
| 14 | GPT-4o Mini | 34.2 | #1667 |
| 15 | GPT-OSS-120B | 33.3 | #704 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=gaeleval-linguistic-competence-gaelic-prompt · How It Works · Data refreshed daily, snapshot 2026-11-21.