Open Japanese LLM - MT: leaderboard
The "MT" subtask of Open Japanese LLM. A task from the Open Japanese LLM Leaderboard, evaluating models on Japanese-language knowledge, reasoning and understanding.
Metric: Score (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 418 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama-3.1-Tulu-3-70B | 86.05 | #482 |
| 2 | Llama-3.3-Swallow-70B-Instruct-v0.4 | 85.99 | #529 |
| 3 | Llama 3.1 70B | 85.56 | #592 |
| 4 | aya-expanse-32B | 85.53 | #692 |
| 5 | Gemma 2 27B (IT) | 85.39 | #645 |
| 6 | Llama 3 70B | 85.31 | #631 |
| 7 | Llama 3.3 70B Instruct | 85.19 | #534 |
| 8 | llm-jp-3.1-13B-instruct4 | 85.17 | #835 |
| 9 | qwen2.5-bakeneko-32B-instruct-v2 | 85.16 | #562 |
| 10 | Llama 3 70B Instruct | 85.09 | #634 |
| 11 | Qwentile2.5-32B-Instruct | 85.06 | #339 |
| 12 | EZO-Qwen2.5-32B-Instruct | 85.01 | #471 |
| 13 | llm-jp-3.1-1.8B-instruct4 | 84.98 | #1273 |
| 14 | Qwen 2.5 32B | 84.96 | #546 |
| 15 | tempmotacilla-cinerea-0308 | 84.96 | #395 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=open-japanese-llm-mt · How It Works · Data refreshed daily, snapshot 2026-10-20.