Thai LLM NLU - xcopa_tha_seacrowd_qa: leaderboard
The Thai LLM Leaderboard, run as an open results dataset by the Thai LLM community. Natural-language understanding accuracy on XCOPA’s Thai causal-alternatives split (SEACrowd harness).
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 70 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o (2024-05-13) | 96.6 | #325 |
| 2 | Qwen 2.5 72B Instruct | 96 | #455 |
| 3 | Claude 3.5 Sonnet (20240620) | 94.4 | #298 |
| 4 | QwQ 32B-Preview | 93.2 | #650 |
| 5 | Sailor2-20B-Chat | 92.8 | #590 |
| 6 | Llama 3.1 70B Instruct | 92.6 | #561 |
| 7 | Llama 3.3 70B Instruct | 92.4 | #534 |
| 8 | GPT-4o Mini (2024-07-18) | 92.4 | #557 |
| 9 | Qwen 2.5 32B Instruct | 92.4 | #499 |
| 10 | Llama 3.1 Nemotron 70B Instruct | 92.4 | #553 |
| 11 | Llama 3.1 70B Cpt Sea Lionv3 Instruct | 92.4 | #548 |
| 12 | Sailor2-8B-Chat | 91.2 | #782 |
| 13 | Qwen 2 72B Instruct | 91 | #577 |
| 14 | DeepSeek R1 Distill Qwen 32B | 90.4 | #655 |
| 15 | Qwen 2.5 14B Instruct | 89.6 | #646 |
No result here: #4 Claude Opus 5.5, #6 GPT-6 Astra, #7 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=thai-llm-nlu-xcopa-tha-seacrowd-qa · How It Works · Data refreshed daily, snapshot 2026-10-21.