Thai LLM NLG - flores200_eng_Latn_tha_Thai_seacrowd_t2t (SacreBLEU): leaderboard
The Thai LLM Leaderboard, run as an open results dataset by the Thai LLM community. Generation quality on FLORES-200 translation from English into Thai, scored by SacreBLEU, the reproducible tokenisation-independent variant (SEACrowd harness).
Metric: SacreBLEU. Source: huggingface.co. Saturation forecast: Estimated already saturated. 68 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude 3.5 Sonnet (20240620) | 46.65 | #299 |
| 2 | GPT-4o (2024-05-13) | 46.21 | #325 |
| 3 | GPT-4o Mini (2024-07-18) | 42.02 | #558 |
| 4 | Llama 3.1 70B Cpt Sea Lionv3 Instruct | 39.87 | #548 |
| 5 | gemma2-9B-cpt-sea-lionv3-instruct | 38.66 | #705 |
| 6 | Llama 3.1 Nemotron 70B Instruct | 38.58 | #554 |
| 7 | Llama 3.1 70B Instruct | 38.57 | #561 |
| 8 | Llama 3 8B Cpt Sea Lionv2 Instruct | 37.8 | #1188 |
| 9 | Llama 3.3 70B Instruct | 37.63 | #534 |
| 10 | DeepSeek R1 Distill Llama 70B | 36.4 | #594 |
| 11 | Llama 3 8B Cpt Sea Lionv2.1 Instruct | 36.3 | #1024 |
| 12 | Qwen 2.5 72B Instruct | 35.58 | #454 |
| 13 | Llama 3.1 8B CPT SEA LIONv3 Instruct | 35.39 | #810 |
| 14 | DeepSeek R1 Distill Qwen 32B | 35.37 | #655 |
| 15 | Gemma 2 9B (IT) | 35.01 | #788 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=thai-llm-nlg-flores200-eng-latn-tha-thai-seacrowd-t2t-sacrebleu · How It Works · Data refreshed daily, snapshot 2026-10-20.