A compact but capable text model built on Qwen3's 8B architecture, this variant carries a 'GDN primed' designation whose specific tuning details aren't publicly documented. It handles a generous 131K token context window, making it comfortable with long documents despite its smaller parameter count. As an open-weight release under Apache 2.0, it's freely inspectable and modifiable, though the nature of any additional fine-tuning beyond the base Qwen3 Instruct training remains unclear.