Open Japanese LLM - ALT E TO J Comet Wmt22: leaderboard
The "ALT E TO J Comet Wmt22" 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 | 91.67 | #692 |
| 2 | Llama-3.3-Swallow-70B-Instruct-v0.4 | 91.61 | #529 |
| 3 | Llama-3.1-Tulu-3-70B | 91.45 | #482 |
| 4 | Gemma 2 27B (IT) | 91.26 | #645 |
| 5 | Llama 3.1 70B | 91.1 | #592 |
| 6 | DeepSeek-R1-Distill-Qwen-32B-Japanese | 91.01 | #638 |
| 7 | tempmotacilla-cinerea-0308 | 91.01 | #395 |
| 8 | Qwentile2.5-32B-Instruct | 91 | #339 |
| 9 | Llama 3.1 Nemotron 70B Instruct | 90.99 | #554 |
| 10 | aya-23-35B | 90.97 | #883 |
| 11 | Qwen 2.5 32B | 90.97 | #546 |
| 12 | aya-expanse-8B | 90.96 | #931 |
| 13 | qwen2.5-bakeneko-32B-instruct-v2 | 90.96 | #562 |
| 14 | Llama 3.3 70B Instruct | 90.95 | #534 |
| 15 | Llama 3.1 70B Instruct | 90.95 | #561 |
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-e-to-j-comet-wmt22 · How It Works · Data refreshed daily, snapshot 2026-10-20.