deep-analysis-research/D2IL-Japanese-Qwen2.5-32B-Instruct-v0.1

TEXT GENERATIONPricing:Input $2.72 / Output $4.8Concurrent Unit Cost:2Model Size:32.8BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Jun 26, 2025License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

D2IL-Japanese-Qwen2.5-32B-Instruct-v0.1 is a 32.5 billion parameter instruction-tuned causal language model from the Qwen2.5 series, developed by Qwen. It features a transformer architecture with RoPE, SwiGLU, and RMSNorm, supporting a context length of 131,072 tokens and generating up to 8,192 tokens. This model significantly enhances coding, mathematics, instruction following, long text generation, and structured data understanding, making it suitable for complex multilingual applications requiring robust performance in Japanese and 28 other languages.

Loading preview...

D2IL-Japanese-Qwen2.5-32B-Instruct-v0.1 Overview

This model is an instruction-tuned variant of the Qwen2.5 series, developed by Qwen, featuring 32.5 billion parameters. It builds upon the Qwen2 architecture with substantial improvements across several key areas. Designed for robust performance, it supports a very long context window of 131,072 tokens for input and can generate responses up to 8,192 tokens.

Key Capabilities

  • Enhanced Knowledge & Reasoning: Significantly improved capabilities in coding and mathematics, leveraging specialized expert models.
  • Instruction Following: Demonstrates stronger adherence to instructions and is more resilient to diverse system prompts, benefiting role-play and chatbot scenarios.
  • Long Text Generation: Excels at generating coherent and relevant long texts, exceeding 8,000 tokens.
  • Structured Data Handling: Improved understanding of structured data like tables and enhanced generation of structured outputs, particularly JSON.
  • Multilingual Support: Provides robust support for over 29 languages, including Japanese, Chinese, English, French, Spanish, and more.
  • Long-Context Processing: Utilizes YaRN for efficient handling of contexts beyond 32,768 tokens, extending up to 128K tokens.

Good For

  • Applications requiring advanced coding and mathematical problem-solving.
  • Chatbots and agents needing precise instruction following and role-play capabilities.
  • Tasks involving extensive text generation or summarization.
  • Processing and generating structured data, such as JSON outputs.
  • Multilingual applications, especially those requiring strong Japanese language understanding and generation.
  • Scenarios demanding very long context understanding, up to 128K tokens.

Finetunes
1 models