SEA LLM Leaderboard: leaderboard
Southeast Asian language leaderboard for SeaExam and SeaBench, evaluating multilingual exam-style knowledge, reasoning, and chat quality in Indonesian, Thai, and Vietnamese.
Metric: Private Average Score (%). Source: huggingface.co. 91 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro (Preview 03-25) | 91.2 | #123 |
| 2 | Claude 3.5 Sonnet (20241022) | 80.8 | #291 |
| 3 | DeepSeek V3 (0324) | 80.8 | #340 |
| 4 | Gemini 1.5 Pro | 79.4 | #445 |
| 5 | Claude 3.7 Sonnet (20250219) | 79.1 | #201 |
| 6 | Gemini 2.0 Flash | 78.4 | #342 |
| 7 | Qwen 2.5 72B | 77.6 | #499 |
| 8 | Llama 4 Maverick Instruct | 76.7 | #456 |
| 9 | Qwen 2.5 72B Instruct | 76.6 | #454 |
| 10 | Gemini 2.0 Flash Lite | 76.4 | #450 |
| 11 | Qwen 2 72B | 75.8 | #521 |
| 12 | GPT-4.1 | 74.6 | #248 |
| 13 | Qwen 2 72B Instruct | 74.5 | #576 |
| 14 | GPT-4o (2024-08-06) | 74.1 | #336 |
| 15 | gemma-3-27B-pt | 73.3 | #502 |
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 · How It Works · Data refreshed daily, snapshot 2026-10-20.