Thai LLM - Reasoning: leaderboard
The Thai LLM Leaderboard, run as an open results dataset by the Thai LLM community. This board is the judge-rated 0-10 score for reasoning in Thai.
Metric: Rating (0-10). Source: huggingface.co. Saturation forecast: Estimated already saturated. 72 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o (2024-05-13) | 9 | #325 |
| 2 | Gemini 1.5 Pro (001) | 7.95 | #444 |
| 3 | Claude 3.5 Sonnet (20240620) | 7.55 | #299 |
| 4 | Qwen 2.5 72B Instruct | 7.15 | #454 |
| 5 | DeepSeek R1 Distill Llama 70B | 7.15 | #594 |
| 6 | DeepSeek R1 Distill Qwen 32B | 6.8 | #655 |
| 7 | GPT-4o Mini (2024-07-18) | 6.7 | #558 |
| 8 | Qwen 2.5 32B Instruct | 6.65 | #498 |
| 9 | Qwen 2.5 14B Instruct | 6.6 | #647 |
| 10 | Llama 3.3 70B Instruct | 6.5 | #534 |
| 11 | Qwen 2 72B Instruct | 6.5 | #576 |
| 12 | Llama 3.1 Nemotron 70B Instruct | 6.45 | #554 |
| 13 | Llama 3.1 70B Cpt Sea Lionv3 Instruct | 6.35 | #548 |
| 14 | Gemma 2 9B (IT) | 5.9 | #788 |
| 15 | Llama 3.1 70B Instruct | 5.85 | #561 |
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-reasoning · How It Works · Data refreshed daily, snapshot 2026-10-20.