Jais-3-0.9B: benchmark results
Provider: Inception (G42), Cerebras & MBZUAI. Released 2026-10-15. Access: Open.
Aggregate Rating 1498 ± 42, rank #2385 of 10916 rated models, from 28 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Jais 3 Model Cards - SimpleQA Verified (Arabic) - Non-hallucination Rate | 79.3 | Non-hallucination rate (%) | 89.3 |
| Jais 3 Model Cards - Arabic IFBench (IFM) | 91.1 | Prompt-level strict accuracy (%) | 85.7 |
| Jais 3 Model Cards - QADI Dialect Identification | 65.9 | Accuracy (%) | 78.6 |
| Jais 3 Model Cards - Arabic Linguistics - Diacritization (IFM) | 83.6 | Accuracy (%) | 53.6 |
| Jais 3 Model Cards - Arabic Linguistics - Parsing (IFM) | 87.4 | Accuracy (%) | 50 |
| Jais 3 Model Cards - QIAS-1 (IFM-audited) | 61.1 | Accuracy (%) | 48.2 |
| Jais 3 Model Cards - Alyah (IFM-audited) | 73.8 | Accuracy (%) | 46.4 |
| Jais 3 Model Cards - Arabic IFEval | 71.3 | Accuracy (%) | 46.4 |
| Jais 3 Model Cards - Arabic Cuisine (IFM) | 73.4 | Accuracy (%) | 42.9 |
| Jais 3 Model Cards - AraDiCE-Culture (IFM-audited) | 72.6 | Accuracy (%) | 39.3 |
| Jais 3 Model Cards - Arabic MMLU with Diacritics | 65.1 | Accuracy (%) | 39.3 |
| Jais 3 Model Cards - Belebele | 81.6 | Accuracy (%) | 39.3 |
Interactive version: theaggregate.ai/model?slug=jais-3-0-9b · How It Works · Data refreshed daily, snapshot 2026-11-08.