Open Japanese LLM - QA — leaderboard
Metric: Score (%). Source: huggingface.co. 418 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | aya-expanse-32B | 64.95 |
| 2 | Llama 3.1 70B | 64.87 |
| 3 | Llama 3 70B | 64.65 |
| 4 | Llama 3.1 70B Instruct | 64.56 |
| 5 | Llama 3.1 Nemotron 70B Instruct | 64.53 |
| 6 | Llama 3.3 70B Instruct | 64.12 |
| 7 | Llama 3 70B Instruct | 61.25 |
| 8 | aya-23-35B | 59.91 |
| 9 | Gemma 2 27B (IT) | 59.32 |
| 10 | Qwentile2.5-32B-Instruct | 58.26 |
| 11 | QwQ 32B-Preview | 57.75 |
| 12 | Llama 2 70B Base | 57.66 |
| 13 | DeepSeek R1 Distill Llama 70B | 56.66 |
| 14 | Qwen 2.5 32B | 56.6 |
| 15 | tempmotacilla-cinerea-0308 | 54.81 |
Interactive version: theaggregate.ai/benchmark?slug=open-japanese-llm-qa · How the rankings work · Data refreshed daily, snapshot 2026-07-22.