Thai LLM NLG - flores200_tha_Thai_eng_Latn_seacrowd_t2t (BLEU): leaderboard

The Thai LLM Leaderboard, run as an open results dataset by the Thai LLM community. Generation quality on FLORES-200 translation from Thai into English, scored by BLEU (SEACrowd harness).

Metric: BLEU. Source: huggingface.co. Saturation forecast: Estimated already saturated. 68 models tracked.

Top models

#ModelScoreOverall rank
1Claude 3.5 Sonnet (20240620)54.81#308
2GPT-4o (2024-05-13)52.96#337
3Llama 3.1 70B Instruct52.76#574
4Llama 3.1 70B Cpt Sea Lionv3 Instruct52.38#559
5Qwen 2.5 72B Instruct52.01#470
6Llama 3 8B Cpt Sea Lionv2 Instruct50.77#1174
7GPT-4o Mini (2024-07-18)50.45#567
8DeepSeek R1 Distill Llama 70B49.94#615
9Gemma 2 9B (IT)49.88#794
10Llama 3.3 70B Instruct49.75#541
11Llama 3 8B Cpt Sea Lionv2.1 Instruct49.59#1011
12Qwen 2.5 32B Instruct49.55#511
13Llama 3.1 8B Instruct49.53#1019
14Llama 3.1 8B CPT SEA LIONv3 Instruct49.3#817
15DeepSeek R1 Distill Qwen 32B48.34#662

No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #6 GPT-6.1 Sol.

Interactive version: theaggregate.ai/benchmark?slug=thai-llm-nlg-flores200-tha-thai-eng-latn-seacrowd-t2t-bleu · How It Works · Data refreshed daily, snapshot 2026-11-06.