Skip to content
AI IntelligenceAug 23, 2026AI Intelligence
Article

SenseTime open-sources 8B multimodal model with native 4K image output

SenseTime open-sourced SenseNova U1.5 Lite, an 8-billion-parameter multimodal model that unifies visual understanding, image generation and editing. It supports native 4K image output, handles constraints on subjects, counts, spatial relationships, text, layouts and visual styles, and improves identity preservation.

Frontier EditorialSource: TechNode
01

Source Brief

SenseTime open-sources 8B multimodal model with native 4K image output: SenseTime open-sourced SenseNova U1.5 Lite, an 8-billion-parameter multimodal model that unifies visual understanding, image generation and editing. It supports native 4K image output, handles constraints on subjects, counts, spatial relationships, text, layouts and visual styles, and improves identity preservation.

02