Sailor2-20B-Chat: benchmark results

Provider: Sea AI Lab. Released 2025-02-18. Access: Open.

Unified ELO 1561 ± 1, rank #443 of 1553 rated models, from 29 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Thai LLM - Coding7.82Rating (0-10)93.7
Thai LLM NLU - xcopa_tha_seacrowd_qa92.8Accuracy (%)92.8
Thai LLM - Social Science8.4Rating (0-10)85.2
Thai LLM - Roleplay7.6Rating (0-10)83.8
Thai LLM - Math6.7Rating (0-10)83.1
Thai LLM - STEM7Rating (0-10)83.1
Thai LLM MC - m3exam_tha_seacrowd_qa60.89Accuracy (%)82.6
Thai LLM NLU - belebele_tha_thai_seacrowd_qa85.78Accuracy (%)82.6
Thai LLM - Writing7.2Rating (0-10)82.4
SEA LLM Leaderboard - SeaExam73.1Private Average Score (%)82.2
Thai LLM MC - thaiexam_qa58.05Accuracy (%)78.3
Thai LLM - Knowledge III4.45Rating (0-10)71.8

Interactive version: theaggregate.ai/model?slug=sailor2-20b-chat · How It Works · Data refreshed daily, snapshot 2026-09-08.