ArtECulture - Arabic: leaderboard
Prediction of the emotion an artwork evokes in the Arabic-speaking culture in ArtECulture (2026), zero-shot. Accuracy in percent, higher is better.
Metric: Accuracy (%; Arabic-speaking culture labels; zero-shot prediction of the majority emotion (nine categories) that the target culture perceives in each of the 679 test artworks; original models without the authors retrieval knowledge or fine-tuning). Source: arxiv.org. Saturation forecast: Around 2033. 16 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.5 27B | 41.53 | #358 |
| 2 | Qwen 3.6 35B A3B | 39.76 | #461 |
| 3 | Gemini 3.5 Flash | 39.62 | #98 |
| 4 | Gemini 3 Flash | 39.18 | #159 |
| 5 | Gemma 4 31B (IT) | 38.44 | #343 |
| 6 | Qwen 3.6 27B | 38 | #339 |
| 7 | Claude Sonnet 4.6 | 35.94 | #158 |
| 8 | Claude Opus 4.8 | 33.04 | #56 |
| 9 | GPT-5.5 | 31.96 | #48 |
| 10 | Qwen 3.5 9B | 31.22 | #803 |
| 11 | Gemma 3 12B (IT) | 28.13 | #1849 |
| 12 | Pixtral-12B | 27.69 | #2493 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=arteculture-arabic · How It Works · Data refreshed daily, snapshot 2026-11-12.