%20--%3e%3cdefs%3e%3cstyle%3e%20.st0%20{%20fill:%20%23061b40;%20}%20.st1%20{%20fill:%20%23306af1;%20}%20.st2%20{%20fill:%20%235ce5cf;%20}%20%3c/style%3e%3c/defs%3e%3cg%3e%3cpath%20class='st0'%20d='M55,10.5h9v3h-9v9h12V7.5h-12v3ZM64,19.5h-6v-3h6v3Z'/%3e%3cpolygon%20class='st0'%20points='69%2016.5%2078%2016.5%2078%2019.5%2069%2019.5%2069%2022.5%2081%2022.5%2081%2013.5%2072%2013.5%2072%2010.5%2081%2010.5%2081%207.5%2069%207.5%2069%2016.5'/%3e%3cpolygon%20class='st0'%20points='95%2010.5%2095%207.5%2083%207.5%2083%2022.5%2095%2022.5%2095%2019.5%2086%2019.5%2086%2016.5%2095%2016.5%2095%2013.5%2086%2013.5%2086%2010.5%2095%2010.5'/%3e%3cpath%20class='st0'%20d='M40,1.5v21h11.6l1.4-1.4v-7.6h0c0,0-1.4-1.5-1.4-1.5l1.4-1.4V2.9l-1.4-1.4h-11.6ZM50,19.5h-7v-6h7v6ZM50,10.5h-7v-6h7v6Z'/%3e%3c/g%3e%3cpath%20class='st1'%20d='M23.1,24L14.7,4.8l-4.9,11.2h3.8l-1.8,4H3.3L12.1,0H2C.9,0,0,.9,0,2v20c0,1.1.9,2,2,2h21.1Z'/%3e%3cpath%20class='st2'%20d='M34,0h-16.8l10.6,24h6.2c1.1,0,2-.9,2-2V2C36,.9,35.1,0,34,0ZM32.5,20h-4V4h4v16Z'/%3e%3c/svg%3e)
SenseTime has brought its lightweight native unified multimodal model to the center of the open-source stage - SenseNova U1.5-Lite-Preview is an early preview version, based on the native NEO-Unify architecture, which is naturally capable of understanding, reasoning, generating, and editing multiple modalities. The most impressive part is that it only uses 8B-MoT parameters but delivers generation and editing quality comparable to commercial closed-source models, effectively using a small but powerful gun to withstand the firepower of the large model arms race.
The upgrades in this preview version are quite substantial. The resolution has been increased to 4K, with more detailed visuals and fewer visual artifacts; the ability to render Chinese and English text and complex layout compositions has significantly improved, making tasks like posters and infographics much easier; the visual control is more precise, no longer causing confusion when dealing with long texts and multiple constraint conditions; native image editing is also more reliable, including style transfer guided by reference images, multi-reference image synthesis, local infographic editing, and text editing.
Compared to the first generation U1, it has taken a significant lead in several generation and editing benchmarks: Qwen-Image-Bench has risen from 47.14 to 55.20 (using PE), ImgEdit-Bench has increased from 3.90 to 4.37, GEdit-Bench-en has climbed from 7.47 to 8.17, and GEdit-Bench-zh has moved up from 7.42 to 8.05. From U1 to U1.5, SenseTime has turned this open-source journey into a footnote of continuous innovation at the foundation model level.
What's even more exciting is still to come: the production-level model for creators, U1Pro, will also be opened for public beta testing soon. This means that SenseTime has already scheduled both "lightweight preview" and "production-ready" ends on the timeline for the unified multimodal line - small models lower the threshold, while large models then meet serious creative demands.
Project address: https://github.com/OpenSenseNova/SenseNova-U1/blob/feat/u1.5_preview/docs/u1.5_preview.md