Flux-inc's Flux-Japanese-Qwen2.5-32B-Instruct-V1.0 is a 32.8 billion parameter instruction-tuned model based on the Qwen2.5-32B-Instruct architecture, specifically optimized for Japanese language tasks. It demonstrates strong performance in Japanese knowledge, reasoning, and language, achieving a Rank-1 position on the Open LLM Japanese LLM Leaderboard with an average score of 0.7417. This model excels particularly in Japanese fundamental analysis, summarization, and code generation, while maintaining general capabilities comparable to its base model.
No reviews yet. Be the first to review!
Hugging Face