A mid-sized multimodal model that handles both text and image inputs, producing text output. As a community-published variant in the Qwen3.8 family, it carries the base architecture's capabilities while being packaged for specific hardware contexts — the '3090' in its name suggests optimization targeting that GPU. Details about any fine-tuning or behavioral modifications applied by the publisher are limited.