llama-3-typhoon-v1.5-8B-instruct: benchmark results
SCB10X's Llama-3-Typhoon-1.5-8B-Instruct, a Thai-English instruction-tuned large language model built on Meta's Llama 3 8B architecture. Provider: SCB 10X. Released 2024-05-06. Access: Open.
Aggregate Rating 1408 ± 19, rank #1198 of 1656 rated models, from 41 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Thai LLM NLG - xl_sum_tha_seacrowd_t2t (ROUGE1) | 29.71 | ROUGE1 | 79.1 |
| Thai LLM NLG - xl_sum_tha_seacrowd_t2t (ROUGE2) | 10.4 | ROUGE2 | 77.6 |
| Thai LLM NLG - iapp_squad_seacrowd_qa (ROUGE1) | 56.66 | ROUGE1 | 76.1 |
| Thai LLM NLG - iapp_squad_seacrowd_qa (ROUGE2) | 44.22 | ROUGE2 | 76.1 |
| Thai LLM NLG - iapp_squad_seacrowd_qa (ROUGEL) | 55.99 | ROUGEL | 76.1 |
| Thai LLM NLG - xl_sum_tha_seacrowd_t2t (ROUGEL) | 19.73 | ROUGEL | 74.6 |
| Open LLM Leaderboard v1 - MMLU | 64.46 | Accuracy (%) (5-shot) | 70.8 |
| Open LLM Leaderboard v1 - GSM8K | 57.16 | Accuracy (%) (5-shot) | 69 |
| Thai LLM - Writing | 6.6 | Rating (0-10) | 63.4 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 53.25 | MC2 (%) (0-shot) | 58.9 |
| Open LLM Leaderboard v1 - WinoGrande | 77.66 | Accuracy (%) (5-shot) | 57.6 |
| Thai LLM - STEM | 6.05 | Rating (0-10) | 56.3 |
Interactive version: theaggregate.ai/model?slug=llama-3-typhoon-v1-5-8b-instruct · How It Works · Data refreshed daily, snapshot 2026-10-20.