SenseTime open-sources 8B unified multimodal SenseNova U1.5 Lite with native 4K generation and precise editing
HeyAmit_ · x · 2026-08-21
SenseTime has open-sourced SenseNova U1.5 Lite, an 8B-parameter lightweight unified multimodal model that jointly learns visual understanding, spatial reasoning, text, generation, and editing. Highlights:
- Native 4K generation with better detail, composition, and realism
- Complex instruction following across subjects, counts, text, layouts, and styles
- Enhanced Chinese/English text rendering and multi-text layout
- Precise image editing via visual markers, bounding boxes, and multi-image references — e.g., replacing a selected infographic element while preserving its position, surrounding layout, and visual relationships
It reportedly outperforms same-size models in instruction following and edit preservation while competing with large commercial models in text rendering. The SenseNova-U1 repo has 5.1k stars on GitHub.
More from Multimodal
- MLP-Splatting (ECCV 2026) code released: MLPs as light-field primitives instead of Gaussians — AjdDavison · 2026-10-16
- Digital human Fae codes her own facial expressions frame-by-frame with Claude Opus 5.5 and FACS — repligate · 2026-10-16
- Meta's Omnilingual ASR, covering 1,600 languages, named a TIME Best Invention of 2026 — AIatMeta · 2026-10-16
- ATAS: Zhejiang-led team builds world-state memory framework for consistent agentic video generation — CSProfKGD · 2026-10-16
- Flux 3 Klein's Default Templates Are Broken Without Its Prompt Upscaler — StrictMove5839 · 2026-10-16
- Kroma Touted as the Most Well-Rounded Image Model Yet on Civitai — Different_Fix_2217 · 2026-10-16