Jais-3-375B-A23B: benchmark results
Provider: Inception (G42), Cerebras & MBZUAI. Released 2026-10-15. Access: Open.
Aggregate Rating 1736 ± 39, rank #218 of 10916 rated models, from 38 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Jais 3 Model Cards - AraGen v3 - Conciseness | 31.5 | 3C3H Conciseness (%) | 100 |
| Jais 3 Model Cards - Arabic IFBench (IFM) | 93.9 | Prompt-level strict accuracy (%) | 100 |
| Jais 3 Model Cards - FannOrFlop MCQ | 73.2 | Accuracy (%) | 100 |
| Jais 3 Model Cards - PaLM-X 2025-2 (IFM-audited) | 95.3 | Accuracy (%) | 100 |
| Jais 3 Model Cards - QADI Dialect Identification | 71.1 | Accuracy (%) | 100 |
| Jais 3 Model Cards - Terminal-Bench 2.1 (Avg@3, Jais Runs) | 73.8 | Accuracy (%) | 100 |
| Jais 3 Model Cards - Arabic Cuisine (IFM) | 89.8 | Accuracy (%) | 98.2 |
| Jais 3 Model Cards - Alyah (IFM-audited) | 87.6 | Accuracy (%) | 96.4 |
| Jais 3 Model Cards - Arabic Linguistics - Parsing (IFM) | 97.7 | Accuracy (%) | 96.4 |
| Jais 3 Model Cards - Belebele | 95.6 | Accuracy (%) | 96.4 |
| Jais 3 Model Cards - HalluScore | 70.9 | Accuracy (%) | 96.4 |
| Jais 3 Model Cards - QIAS-2 (IFM-audited) | 87.7 | Accuracy (%) | 96.4 |
Interactive version: theaggregate.ai/model?slug=jais-3-375b-a23b · How It Works · Data refreshed daily, snapshot 2026-11-08.