SeaEval - Cultural Reasoning - PH-Eval (Zero-Shot): leaderboard
SeaEval zero-shot PH-Eval evaluation measuring culturally grounded reasoning and regional knowledge for Southeast Asian and comparison contexts.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 25 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 2.5 72B Instruct | 72 | #454 |
| 2 | Qwen 2.5 32B Instruct | 70 | #498 |
| 3 | Llama 3.1 70B Cpt Sea Lionv3 Instruct | 69 | #548 |
| 4 | Llama 3.1 70B Instruct | 68 | #561 |
| 5 | Llama 3 70B Instruct | 63 | #634 |
| 6 | Qwen 2 72B Instruct | 62 | #576 |
| 7 | Qwen 2.5 14B Instruct | 60 | #647 |
| 8 | Llama 3.1 8B CPT SEA LIONv3 Instruct | 60 | #810 |
| 9 | gemma2-9B-cpt-sea-lionv3-instruct | 59 | #705 |
| 10 | Gemma 2 9B (IT) | 58 | #788 |
| 11 | Llama 3 8B Instruct | 58 | #1129 |
| 12 | Llama 3.1 8B Instruct | 57 | #1035 |
| 13 | Llama 3 8B Cpt Sea Lionv2.1 Instruct | 56 | #1024 |
| 14 | Llama 3 8B Cpt Sea Lionv2 Instruct | 56 | #1188 |
| 15 | Qwen 2.5 7B Instruct | 55 | #862 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=seaeval-cultural-reasoning-ph-eval-zero-shot · How It Works · Data refreshed daily, snapshot 2026-10-20.