ALBA - Culture-bound Semantics: leaderboard
Metric: Mean judge score (0-100) on ALBA's 100 expert-written pt-PT culture-bound semantics questions (the meaning and pt-PT equivalents of idioms, sayings and proverbs), judged by Gemini-2.5-Pro with an expert pt-PT rubric (precision, linguistic quality, completeness), three expert-rated few-shot examples and greedy decoding, 1-5 rating rescaled to 0-100; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 15 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 88.3 | #91 |
| 2 | Gemini 2.5 Pro | 85.3 | #145 |
| 3 | Gemma 3 12B (IT) | 41.5 | #655 |
| 4 | EuroLLM-9B-Instruct | 40 | #1043 |
| 5 | Apertus-8B-Instruct-2509 | 38.3 | #1048 |
| 6 | Ministral-8B-Instruct-2410 | 37.5 | #1080 |
| 7 | Qwen 3 8B | 35.8 | #667 |
| 8 | Gemma 2 9B (IT) | 35.5 | #774 |
| 9 | salamandra-7B-instruct | 30.5 | #1390 |
| 10 | Qwen 2.5 7B Instruct | 24 | #846 |
| 11 | Llama 3.1 8B Instruct | 23.5 | #1018 |
| 12 | Mistral 7B Instruct | 17.5 | #1345 |
| 13 | OLMo-2-1124-7B-Instruct | 9.5 | #1349 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=alba-culture-bound-semantics · How It Works · Data refreshed daily, snapshot 2026-10-11.