Thai LLM - Writing Leaderboard

Thai LLM - Writing: leaderboard

Metric: Rating (0-10). Source: huggingface.co. 72 models tracked.

#	Model	Score
1	GPT-4o Mini (2024-07-18)	8.35
2	GPT-4o (2024-05-13)	8.15
3	Gemini 1.5 Flash (001)	8
4	Gemini 1.5 Pro (001)	7.9
5	gemma2-9B-cpt-sea-lionv3-instruct	7.8
6	Claude 3.5 Sonnet (20240620)	7.6
7	Llama 3.1 70B Cpt Sea Lionv3 Instruct	7.55
8	Llama 3 8B Cpt Sea Lionv2.1 Instruct	7.5
9	Sailor2-8B-Chat	7.45
10	Llama 3.1 8B Cpt Sea Lionv3 Instruct	7.45
11	DeepSeek R1 Distill Qwen 32B	7.3
12	Qwen 2.5 72B Instruct	7.2
13	Qwen 2.5 14B Instruct	7.05
14	DeepSeek R1 Distill Llama 70B	7.05
15	Llama 3.1 Nemotron 70B Instruct	7.05

Interactive version: theaggregate.ai/benchmark?slug=thai-llm-writing · How It Works · Data refreshed daily, snapshot 2026-09-05.