Harsha901/Qwen3-4B-Inst-Math-Reasoning-SFT
TEXT GENERATIONConcurrency Cost:1Model Size:4BQuant:BF16Ctx Length:32kPublished:Dec 8, 2025License:apache-2.0Architecture:Transformer Open Weights Cold

Harsha901/Qwen3-4B-Inst-Math-Reasoning-SFT is a 4 billion parameter Qwen3-Instruct model, fine-tuned by Harsha901 for enhanced mathematical reasoning and step-by-step problem-solving. Utilizing Unsloth for accelerated training, this model excels at generating logically structured reasoning chains for math problems, algebra, and word problems. It is optimized for educational assistants and reasoning benchmarks, providing improved instruction adherence and more stable reasoning outputs.

Loading preview...