SeaEval - Cultural Reasoning - SG-Eval v1 Cleaned (Zero-Shot) — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 22 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 2.5 72B Instruct | 72.06 |
| 2 | Qwen 2.5 14B Instruct | 69.12 |
| 3 | Qwen 2 72B Instruct | 67.65 |
| 4 | Llama 3.1 70B Instruct | 66.18 |
| 5 | Llama 3 70B Instruct | 66.18 |
| 6 | Llama 3 8B Cpt Sea Lionv2.1 Instruct | 66.18 |
| 7 | Qwen 2.5 32B Instruct | 64.71 |
| 8 | Qwen 2 7B Instruct | 63.24 |
| 9 | Gemma 2 9B (IT) | 60.29 |
| 10 | Llama 3 8B Instruct | 58.82 |
| 11 | Qwen 2.5 7B Instruct | 58.82 |
| 12 | Qwen 2.5 3B Instruct | 58.82 |
| 13 | SeaLLMs-v3-7B-Chat | 58.82 |
| 14 | Llama 3.1 8B Instruct | 57.35 |
| 15 | Sailor2-8B-Chat | 57.35 |
Interactive version: theaggregate.ai/benchmark?slug=seaeval-cultural-reasoning-sg-eval-v1-cleaned-zero-shot · How the rankings work · Data refreshed daily, snapshot 2026-07-22.