Alelcv27/Qwen3-4B-INST-Math-Code

TEXT GENERATIONConcurrency Cost:1Model Size:4BQuant:BF16Ctx Length:32kTool Calling:SupportedPublished:Jun 4, 2026License:apache-2.0Architecture:Transformer Open Weights Cold

Alelcv27/Qwen3-4B-INST-Math-Code is a 4 billion parameter Qwen3 instruction-tuned model developed by Alelcv27, specifically fine-tuned for mathematical and coding tasks. This model leverages Unsloth and Huggingface's TRL library for accelerated training, making it efficient for specialized applications. With a 32768 token context length, it is designed to excel in complex problem-solving within its target domains.

Loading preview...

Overview

Alelcv27/Qwen3-4B-INST-Math-Code is a specialized 4 billion parameter instruction-tuned model based on the Qwen3 architecture. Developed by Alelcv27, this model is a further fine-tuned version of Alelcv27/Qwen3-4B-INST-Math, focusing on enhancing its capabilities in both mathematical reasoning and code generation.

Key Characteristics

  • Base Model: Qwen3 architecture.
  • Parameter Count: 4 billion parameters.
  • Context Length: Supports a substantial 32768 token context window, beneficial for handling longer problem descriptions or code snippets.
  • Training Efficiency: The model was trained significantly faster using Unsloth and Huggingface's TRL library, indicating an optimized training process.
  • Specialized Fine-tuning: It is specifically fine-tuned for mathematical and coding tasks, building upon its predecessor's math capabilities.

Good For

  • Mathematical Problem Solving: Excels in tasks requiring mathematical reasoning and computation.
  • Code Generation: Suited for generating and understanding code across various programming languages.
  • Efficient Deployment: Its optimized training suggests potential for efficient inference, making it suitable for applications where speed is a factor.
  • Specialized AI Applications: Ideal for developers building applications that require strong performance in both math and code domains.

Finetunes
1 models