Jais-3-3.7B: benchmark results
Provider: Inception (G42), Cerebras & MBZUAI. Released 2026-10-15. Access: Open.
Aggregate Rating 1625 ± 40, rank #791 of 10916 rated models, from 35 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Jais 3 Model Cards - APEX-Agents (Pass@1) | 16.7 | Pass@1 (%) | 100 |
| Jais 3 Model Cards - WildClawBench (Mean Score@3) | 40.9 | Mean Score@3 | 100 |
| Jais 3 Model Cards - Arabic IFBench (IFM) | 92.6 | Prompt-level strict accuracy (%) | 96.4 |
| Jais 3 Model Cards - SimpleQA Verified (Arabic) - Non-hallucination Rate | 84.2 | Non-hallucination rate (%) | 96.4 |
| Jais 3 Model Cards - Arabic Linguistics - Parsing (IFM) | 97.3 | Accuracy (%) | 87.5 |
| Jais 3 Model Cards - BrowseComp-Plus | 64.8 | Accuracy (%) | 83.3 |
| Jais 3 Model Cards - QADI Dialect Identification | 68.8 | Accuracy (%) | 82.1 |
| Jais 3 Model Cards - Arabic IFEval | 79 | Accuracy (%) | 78.6 |
| Jais 3 Model Cards - GAIA v2 lit (Arabic) | 45.5 | Pass@1 (%) | 76.1 |
| Jais 3 Model Cards - IslamicEval 2025-1B (IFM-audited) | 74.5 | Accuracy (%) | 71.4 |
| Jais 3 Model Cards - Alyah (IFM-audited) | 79.7 | Accuracy (%) | 67.9 |
| Jais 3 Model Cards - QIAS-1 (IFM-audited) | 83.3 | Accuracy (%) | 67.9 |
Interactive version: theaggregate.ai/model?slug=jais-3-3-7b · How It Works · Data refreshed daily, snapshot 2026-11-08.