SEA LLM Leaderboard - SeaExam - Thai (Public): leaderboard
SeaExam public set (the downloadable SeaLLMs dataset) accuracy in Thai: multiple-choice questions from real human exams testing world knowledge and reasoning, 3-shot for open models and zero-shot for API models. 0-100, higher is better; from DAMO-NLP-SG's LLM Leaderboard for SEA.
Metric: Public Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 91 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro (Preview 03-25) | 92.9 | #126 |
| 2 | Claude 3.5 Sonnet (20241022) | 80.9 | #302 |
| 3 | Claude 3.7 Sonnet (20250219) | 80.9 | #209 |
| 4 | Gemini 2.0 Flash | 79 | #351 |
| 5 | GPT-4.1 | 78 | #258 |
| 6 | Gemini 1.5 Pro | 78 | #454 |
| 7 | DeepSeek V3 (0324) | 77.8 | #349 |
| 8 | GPT-4o (2024-08-06) | 77 | #343 |
| 9 | Qwen 2.5 72B | 76.3 | #512 |
| 10 | Llama 4 Maverick Instruct | 75.8 | #467 |
| 11 | Qwen 2.5 72B Instruct | 75.2 | #469 |
| 12 | Qwen 2 72B | 74.4 | #515 |
| 13 | Gemini 2.0 Flash Lite | 74.2 | #455 |
| 14 | GPT-4.1 Mini | 72.2 | #363 |
| 15 | Qwen 2 72B Instruct | 71.7 | #579 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=sea-llm-leaderboard-seaexam-thai-public · How It Works · Data refreshed daily, snapshot 2026-11-05.