---
license: apache-2.0
---

The Qwen3 model was fine-tuned using the Opus 4.6 reasoning dataset, 
with a focus on enhancing multi-step inference, logical consistency, and 
instruction following across complex tasks.