mastema007/Qwen2.5-3B-Instruct

TEXT GENERATIONPricing:Input $0.32 / Cached $0.064 / Output $1.6Concurrent Unit Cost:1Model Size:3.1BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Sep 14, 2026License:otherArchitecture:Transformer Featherless Exclusive Cold

Qwen2.5-3B-Instruct is a 3.09 billion parameter instruction-tuned causal language model developed by Qwen, part of the latest Qwen2.5 series. It features significantly improved capabilities in coding, mathematics, and instruction following, with a context length of 32,768 tokens. This model excels at generating long texts, understanding structured data, and producing structured outputs like JSON, while also offering multilingual support for over 29 languages.

Loading preview...

Qwen2.5-3B-Instruct: An Enhanced Instruction-Tuned LLM

This repository hosts the instruction-tuned 3.09 billion parameter Qwen2.5 model, part of the latest Qwen large language model series. Developed by Qwen, this model builds upon its predecessors with significant advancements across several key areas.

Key Capabilities & Improvements

  • Enhanced Knowledge & Reasoning: Features substantially improved capabilities in coding and mathematics, leveraging specialized expert models.
  • Superior Instruction Following: Demonstrates significant improvements in adhering to instructions, generating long texts (over 8K tokens), and understanding structured data (e.g., tables).
  • Robust Output Generation: Excels at generating structured outputs, particularly JSON, and is more resilient to diverse system prompts, enhancing role-play and chatbot condition-setting.
  • Extended Context & Multilingual Support: Supports a full context length of 32,768 tokens and offers robust multilingual support for over 29 languages, including Chinese, English, French, Spanish, and more.

Architecture & Features

Qwen2.5-3B-Instruct is a causal language model built on a transformer architecture, incorporating features like RoPE, SwiGLU, RMSNorm, and attention QKV bias. It has 36 layers and 16 attention heads (with 2 for KV in GQA configuration).

Good For

  • Applications requiring strong coding and mathematical reasoning.
  • Tasks involving long-form text generation and complex instruction following.
  • Scenarios needing reliable structured data understanding and JSON output generation.
  • Multilingual applications across a broad range of languages.

Finetunes
1 models