hugging-apps/sensenova-u1-8b-mot-interleaved

SenseNova-U1-8B-MoT-Interleaved

This Space demonstrates SenseNova-U1-8B-MoT-Interleaved, a native any-to-any multimodal model that unifies visual understanding and image generation within a single architecture (NEO-unify).

The model can generate interleaved text and images in a single response — perfect for illustrated tutorials, storybooks, multi-page presentations, and drawing guides.

Usage

Enter a text prompt describing what you want to create. The model will generate text with images woven in. You can optionally provide an input image for image-conditioned generation.

Model

• —Architecture: NEO-unify (Mixture of Transformers) — Qwen3 LLM backbone + Vision Tower + Flow Matching Head
• —Parameters: ~17.5B (bf16)
• —Capabilities: Text-to-image, image-to-text, interleaved text+image generation, image editing