Llama 3 8B Cpt Sahabatai V1 Instruct: benchmark results
GoTo/AI Singapore's Sahabat-AI Llama 3 8B Instruct, an Indonesian-focused chat model continually pretrained from Llama 3 8B and instruction-tuned on Indonesian, Javanese, Sundanese and English data. Provider: Sahabat-AI. Released 2024-11-06. Access: Open.
Aggregate Rating 1465 ± 25, rank #910 of 1656 rated models, from 18 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Sahabat-AI - Indonesian Average | 49.98 | Average Score (%) | 87.5 |
| Sahabat-AI - Javanese (JV) | 49.64 | Score (%) | 87.5 |
| Open LLM Leaderboard - MuSR | 44.88 | Score | 82.8 |
| Sahabat-AI - Indonesian (ID) | 56.56 | Score (%) | 81.2 |
| Sahabat-AI - Sundanese (SU) | 43.74 | Score (%) | 75 |
| Open LLM Leaderboard - IFEval | 52.38 | Score | 62.4 |
| Open LLM Leaderboard - MATH Level 5 | 12.76 | Score | 56.8 |
| Open LLM Leaderboard - MMLU-Pro | 34.53 | Score | 50.1 |
| Open LLM Leaderboard - BBH | 49.51 | Score | 46.3 |
| GoTo-AI Leaderboard - Balinese | 38.8 | Score (0-100) | 22.2 |
| GoTo-AI Leaderboard - Batak Toba | 26.85 | Score (0-100) | 22.2 |
| Open LLM Leaderboard - GPQA | 26.68 | Score | 21.2 |
Interactive version: theaggregate.ai/model?slug=llama-3-8b-cpt-sahabatai-v1-instruct · How It Works · Data refreshed daily, snapshot 2026-10-20.