Jais 3 Model Cards - Belebele: leaderboard
Belebele reading comprehension, as printed in the card's Arabic knowledge section (the card does not name the language split). MBZUAI's Institute of Foundation Models reported this on the Jais 3 model cards (released October 15, 2026), comparing each Jais 3 size with size-matched reference models; this board collects all six cards. The cards state that all reported results use high reasoning effort. Vendor-reported. Metric: Accuracy (%). Higher is better. Source: https://huggingface.co/collections/IFM/jais-3-6abba9df4f1173ec7ce2929d
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 29 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemma 4 31B (IT) | 95.9 | #174 |
| 2 | Qwen3.8-Flash-Next | 95.5 | #69 |
| 3 | Qwen 3.5 397B A17B | 94.8 | #144 |
| 4 | Qwen 3.8 27B | 94.4 | #104 |
| 5 | Muse Glimmer 30B | 94.3 | #178 |
| 6 | Gemma 4 12B (IT) | 93.6 | #371 |
| 7 | Nemotron 3 Ultra | 93.1 | #234 |
| 8 | MiniMax-M3 | 92.3 | #137 |
| 9 | Qwen 3.5 9B | 89.2 | #377 |
| 10 | gemma-4-E4B-it | 88.7 | #566 |
| 11 | Falcon-H1-34B-Instruct | 86.6 | #666 |
| 12 | Qwen 3.5 4B | 84 | #473 |
| 13 | Fanar-1-9B-Instruct | 79.9 | #633 |
| 14 | Gemma 4 E2B | 79.6 | #1076 |
| 15 | Falcon-H1-7B-Instruct | 72.9 | #1043 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=jais-3-model-cards-belebele · How It Works · Data refreshed daily, snapshot 2026-10-19.