SeaEval - Fundamental NLP Tasks - CoLA (Zero-Shot): leaderboard
Metric: Accuracy (%). Source: huggingface.co. 25 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 2.5 72B Instruct | 85.71 |
| 2 | Llama 3.1 70B Instruct | 85.04 |
| 3 | Llama 3.1 70B Cpt Sea Lionv3 Instruct | 84.85 |
| 4 | Qwen 2.5 32B Instruct | 84.28 |
| 5 | Llama 3 70B Instruct | 83.51 |
| 6 | Qwen 2 72B Instruct | 83.41 |
| 7 | gemma2-9B-cpt-sea-lionv3-instruct | 82.55 |
| 8 | Qwen 2.5 14B Instruct | 80.63 |
| 9 | Gemma 2 9B (IT) | 79.39 |
| 10 | Qwen 2.5 7B Instruct | 79.1 |
| 11 | Sailor2-8B-Chat | 79 |
| 12 | Qwen 2 7B Instruct | 78.72 |
| 13 | SeaLLMs-v3-7B-Chat | 78.52 |
| 14 | Llama 3.1 8B Cpt Sea Lionv3 Instruct | 76.99 |
| 15 | Qwen 2.5 1.5B Instruct | 74.98 |
Interactive version: theaggregate.ai/benchmark?slug=seaeval-fundamental-nlp-tasks-cola-zero-shot · How It Works · Data refreshed daily, snapshot 2026-09-05.