NOLLI - Jamo Composition: leaderboard
Metric: Accuracy (%). Source: arxiv.org. 15 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.5 (Medium) | 87.3 |
| 2 | Gemini 3.1 Pro (Preview) (Medium) | 57.3 |
| 3 | Claude Opus 4.8 (Medium) | 47 |
| 4 | Qwen 3.5 397B A17B (Non-reasoning) | 46.3 |
| 5 | DeepSeek V4 Flash | 17.3 |
| 6 | Gemma 4 31B (IT) (Thinking) | 13.7 |
| 7 | Qwen 3.5 9B (Non-reasoning) | 12.3 |
| 8 | GPT-OSS-120B (Medium) | 8.7 |
| 9 | Solar Open 100B | 3.7 |
| 10 | Llama 4 Maverick | 2 |
| 11 | Qwen 3.5 27B | 1.3 |
| 12 | Llama 3.1 8B Instruct | 1 |
| 13 | EXAONE-4.0-32B (Thinking) | 0 |
Interactive version: theaggregate.ai/benchmark?slug=nolli-jamo-composition · How It Works · Data refreshed daily, snapshot 2026-09-19.