SeaEval - Cultural Reasoning - SG-Eval (Zero-Shot) — leaderboard

Metric: Accuracy (%). Source: huggingface.co. 22 models tracked.

Top models

#ModelScore
1Qwen 2.5 72B Instruct76.7
2Qwen 2.5 14B Instruct76.7
3Qwen 2 72B Instruct73.79
4Llama 3.1 70B Instruct71.84
5Qwen 2.5 32B Instruct71.84
6SeaLLMs-v3-7B-Chat71.84
7Llama 3 70B Instruct70.87
8Qwen 2 7B Instruct67.96
9Gemma 2 9B (IT)66.99
10Qwen 2.5 7B Instruct66.99
11Llama 3 8B Cpt Sea Lionv2.1 Instruct66.02
12Llama 3 8B Instruct65.05
13Llama 3.1 8B Instruct64.08
14Qwen 2.5 3B Instruct63.11
15Sailor2-8B-Chat60.19

Interactive version: theaggregate.ai/benchmark?slug=seaeval-cultural-reasoning-sg-eval-zero-shot · How the rankings work · Data refreshed daily, snapshot 2026-07-22.