Saka-1.5B: benchmark results
Sakalti's Saka 1.5B, a TIES merge of their SJT-1.5B-Alpha base with SakanaAI's TinySwallow-1.5B-Instruct. Provider: Sakalti. Released 2025-02-07. Access: Open.
Aggregate Rating 1348 ± 25, rank #1390 of 1656 rated models, from 179 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Arabic LLM - Arabic MMLU HT College Mathematics | 31 | Accuracy (%) | 51.9 |
| Open Arabic LLM - Arabic MMLU HT Abstract Algebra | 28 | Accuracy (%) | 48.5 |
| Open Arabic LLM - Arabic MMLU HT High School Mathematics | 28.52 | Accuracy (%) | 48.5 |
| Open Arabic LLM - Arabic MMLU LAW (Professional) | 62.1 | Accuracy (%) | 48.5 |
| Open Japanese LLM - CG | 5.82 | Score (%) | 46.2 |
| Open LLM Leaderboard - GPQA | 29.03 | Score | 45 |
| Open Japanese LLM - Mbpp Pylint Check | 13.65 | Score (%) | 41.9 |
| Open Arabic LLM - Arabic MMLU Islamic Studies (High School) | 51.8 | Accuracy (%) | 41.4 |
| Open LLM Leaderboard - MATH Level 5 | 8.01 | Score | 40.9 |
| Open Japanese LLM - Niilc Char F1 | 45.17 | Score (%) | 39.7 |
| Open Japanese LLM - Wiki NER SET F1 | 3.54 | Score (%) | 39.4 |
| Open Arabic LLM - Arabic MMLU HT Human Aging | 43.5 | Accuracy (%) | 38.3 |
Interactive version: theaggregate.ai/model?slug=saka-1-5b · How It Works · Data refreshed daily, snapshot 2026-10-20.