Open Japanese LLM - ALT J TO E Bleu EN: leaderboard
The "ALT J TO E Bleu EN" 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 | aya-expanse-32B | 19.43 | #692 |
| 2 | Llama-3.1-Tulu-3-70B | 19.16 | #482 |
| 3 | Llama 3 70B | 18.99 | #631 |
| 4 | Llama-3.3-Swallow-70B-Instruct-v0.4 | 18.86 | #529 |
| 5 | Llama 3.1 70B | 18.7 | #592 |
| 6 | Gemma 2 27B (IT) | 18.52 | #645 |
| 7 | Llama 3.1 70B Instruct | 18.47 | #561 |
| 8 | llm-jp-3.1-13B-instruct4 | 18.45 | #835 |
| 9 | Llama 3.1 Nemotron 70B Instruct | 18.31 | #554 |
| 10 | Llama 3.3 70B Instruct | 18.13 | #534 |
| 11 | Llama 2 70B Base | 18.02 | #1025 |
| 12 | Qwen 2.5 32B | 17.98 | #546 |
| 13 | aya-23-35B | 17.94 | #883 |
| 14 | oxyge1-33B | 17.79 | #690 |
| 15 | Qwen 2.5 32B Instruct | 17.77 | #498 |
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-alt-j-to-e-bleu-en · How It Works · Data refreshed daily, snapshot 2026-10-20.