PLCC - Grammar: leaderboard
Metric: Accuracy (%). Source: huggingface.co. 227 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek V4.1 Flash | 98 |
| 2 | Gemini 3.5 Flash | 97 |
| 3 | GPT-5.5 (High) | 97 |
| 4 | GPT-6 | 95 |
| 5 | Gemini 3.7 Flash | 95 |
| 6 | Gemini 3.8 Flash | 95 |
| 7 | Claude Fable 5.1 | 95 |
| 8 | Claude Fable 5 | 94 |
| 9 | Gemini 3.6 Flash | 94 |
| 10 | Gemini 3.1 Pro (Preview) | 93 |
| 11 | DeepSeek V4 Pro | 92 |
| 12 | Claude Opus 5 | 92 |
| 13 | Inkling | 92 |
| 14 | DeepSeek V4 Flash (0731) | 91 |
| 15 | GPT-5.6 Sol (Medium) | 91 |
Interactive version: theaggregate.ai/benchmark?slug=plcc-grammar · How It Works · Data refreshed daily, snapshot 2026-09-19.