QIMMA - AraDiCE-Culture: leaderboard
Metric: Accuracy (%, log-likelihood multiple choice). Source: huggingface.co. 83 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3.5 397B A17B | 82.78 |
| 2 | Yehia-7B-preview | 81.67 |
| 3 | ALLaM-7B-Instruct-preview | 78.89 |
| 4 | Jais-2-70B-Chat | 78.89 |
| 5 | Llama 3.3 70B Instruct | 77.22 |
| 6 | Qwen 2.5 72B Instruct | 77.22 |
| 7 | Gemma 4 31B | 77.22 |
| 8 | jais-adapted-70B-chat | 76.67 |
| 9 | AceGPT-v2-32B-Chat | 76.67 |
| 10 | jais-adapted-13B-chat | 75.56 |
| 11 | Qwen 3.5 122B A10B | 74.44 |
| 12 | Fanar-1-9B-Instruct | 74.44 |
| 13 | Fanar-1-9B | 74.44 |
| 14 | AIC-1 | 73.33 |
| 15 | Gemma 4 26B A4B | 71.67 |
Interactive version: theaggregate.ai/benchmark?slug=qimma-aradice-culture · How It Works · Data refreshed daily, snapshot 2026-09-19.