SenseNova-U1 8B MoT by SenseNova — 32K Context

8B-parameter end-to-end unified multimodal model based on NEO-Unify architecture. Native unified text and image understanding + generation without separate visual encoder or VAE. Supports visual Q&A, text-to-image generation, image editing, and interleaved text-image generation. Context length up to 32K tokens. Available in both 8B dense MoT and A3B MoE variants. Optimized for infographic generation and complex visual reasoning.