On April 10th at SenseTime’s technology exchange EVEnt, the firm introducedits latest multi-modal large model, SenseNova V6, along with the SenseCore 2.0system. The new iteration is designed to blend text, images, and videos,offering users a more Seamless and diVerse interactive experience.
The SenseNova V6 series consists of four versions. Among them, SenseNova V6Prostands out with its 620 billion-parameter hybrid expert ARChitecture,highlighting formidable multi-modal fusion capabilities. SenseNova V6ReasonerPro bolsters multi-modal reasoning abilities, enabling in-depth logicalanalysis. SenseNova V6Video is centered around video understanding, capable ofsummarizing and conducting in-depth analysis of video content. SenseNovaV6Omni is a compact, full-modal interactive model that combines language,sPeech, and video for real-time interaction.
Demonstrations illustrated the distinctive multi-modal capabilities ofSenseNova V6. Users could interact with the model by presenting photos ofhandwritten math problems. The model not only solved them but also analyzeduser responses, guiding users through the solution process via voice andproviding real-time assistance, giving the impression of a personal tutor.
SenseTime co-founder, Linda Hua, indicated that future interactions willsurely be multi-modal. SenseTime aims to grasp the core technologies for suchinteractions. Hua also pointed out the relative lack of domestic companiesdeveloPing multi-modal reasoning and interaction capabilities. SenseTime hopesto utilize its edge in computer vision to gain an early foothold in the multi-modal large model market.
Moreover, the multi-modal capabilities of SenseNova V6Pro are on par withleading international models such as gemini 2.0Pro and GPT-4.5. SenseTimeemphasizes strong reasoning, strong interaction, and long-term memory as threecrucial technological breakthroughs. These capabilities enable the model tobetter comprehend human intent and facilitate more engaging user interactions.
SenseTime plans to incorporate SenseNova V6 into real-world applicationsacross various sectors, including education, translation, and tourism. Bycollaborating with embodied AI company Fourier, SenseTime aims to endow robotswith enhanced environmental understanding and human-robot interactioncapabilities, truly realizing a more intelligent future.
