llama3.2-typhoon2-3B-instruct: benchmark results

Provider: Meta. Access: Open.

Unified ELO 1441 ± 28, rank #1579 of 2656 rated models, from 36 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Thai LLM NLG - iapp_squad_seacrowd_qa (ROUGE1)72.36ROUGE194
Thai LLM NLG - iapp_squad_seacrowd_qa (ROUGEL)72.07ROUGEL94
Thai LLM - Knowledge III5.15Rating (0-10)90.1
Thai LLM NLG - iapp_squad_seacrowd_qa (ROUGE2)57.26ROUGE289.6
Thai LLM NLG - xl_sum_tha_seacrowd_t2t (ROUGE2)11.13ROUGE286.6
Thai LLM NLG - xl_sum_tha_seacrowd_t2t (ROUGEL)21.65ROUGEL86.6
Thai LLM NLG - xl_sum_tha_seacrowd_t2t (ROUGE1)31.2ROUGE185.1
Thai LLM NLG - flores200_tha_Thai_eng_Latn_seacrowd_t2t (BLEU)49.61BLEU76.1
Thai LLM - Writing6.8Rating (0-10)67.6
Thai LLM NLG - flores200_eng_Latn_tha_Thai_seacrowd_t2t (BLEU)24.32BLEU64.2
Thai LLM NLG - flores200_eng_Latn_tha_Thai_seacrowd_t2t (SacreBLEU)34.86SacreBLEU62.7
Thai LLM NLG - flores200_tha_Thai_eng_Latn_seacrowd_t2t (SacreBLEU)32.23SacreBLEU62.7

Interactive version: theaggregate.ai/model?slug=llama3-2-typhoon2-3b-instruct · How It Works · Data refreshed daily, snapshot 2026-09-19.