FULL STORY

SenseTime Open-Sources SenseNova U1.5, an 8B Unified Multimodal Model

SenseTime released and open-sourced SenseNova U1.5 Lite, an 8B unified multimodal model supporting 4K image generation, dense text rendering and precise editing, drawing continued community attention.

2026-07-31 ~ 2026-08-21 · 4 episodes · 10 posts

Episode 1 · SenseNova U1.5 Lite Supports Native 4K Image Generation (2026-07-31, 2 posts)

SenseNova released the SenseNova U1.5 Lite preview, a multimodal model supporting native 4K image generation and dense text rendering. By improving the image head structure, it reduces grid artifacts and achieves significant progress in multiple image generation and editing benchmarks.

Episode 2 · SenseNova Open-Sources U1.5 Multimodal Model for 4K Image Generation (2026-08-03, 3 posts)

SenseNova has open-sourced SenseNova U1.5-Lite-Preview, an 8B unified multimodal model. Based on the NEO-Unify architecture, the lightweight model integrates language and visual capabilities, natively supporting 4K image generation and high-precision structured editing.

Episode 3 · SenseTime Releases SenseNova U1.5 8B Multimodal Model (2026-08-05, 2 posts)

SenseTime has released and open-sourced SenseNova U1.5-Lite-Preview, an 8B native unified multimodal model based on the NEO-Unify architecture. Despite its lightweight size, it efficiently supports 4K image generation and editing.

Episode 4 · SenseTime Open-Sources 8B Unified Multimodal Model SenseNova U1.5 Lite (2026-08-21, 3 posts)

SenseTime open-sourced SenseNova U1.5 Lite, an 8B unified multimodal model combining perception, generation, and native editing with 4K resolution output in a single path, eliminating the need for complex pipelines.