Open Japanese LLM - Wiki Reading Char F1: leaderboard
The "Wiki Reading Char F1" 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 | llm-jp-3.1-13B-instruct4 | 94.25 | #835 |
| 2 | llm-jp-3.1-1.8B-instruct4 | 91.66 | #1273 |
| 3 | Llama-3.3-Swallow-70B-Instruct-v0.4 | 90.2 | #529 |
| 4 | Llama 3.1 70B | 87.2 | #592 |
| 5 | aya-expanse-32B | 87.05 | #692 |
| 6 | llm-jp-3-13B | 86.93 | #1219 |
| 7 | Llama 3 70B | 86.82 | #631 |
| 8 | Llama 3.1 70B Instruct | 86.16 | #561 |
| 9 | Llama-3.1-Tulu-3-70B | 85.97 | #482 |
| 10 | Llama 3.1 Nemotron 70B Instruct | 85.82 | #554 |
| 11 | Llama 3.3 70B Instruct | 85.01 | #534 |
| 12 | Llama 3 70B Instruct | 83.93 | #634 |
| 13 | qwen2.5-bakeneko-32B-instruct-v2 | 83.46 | #562 |
| 14 | Gemma 2 27B (IT) | 83.03 | #645 |
| 15 | Qwen 2.5 32B Instruct | 82.04 | #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-wiki-reading-char-f1 · How It Works · Data refreshed daily, snapshot 2026-10-20.