SenseChat

SenseChat (商量) is the chat product of SenseTime (商汤), Shanghai's computer-vision company, running on the SenseNova (日日新) system the group sells atop its own SenseCore compute platform. The consumer page we opened bills itself as an “AI生活小帮手” and shows range rather than specs: writing and translation, code help, image generation, multi-file document analysis and everyday Q&A, in a Chinese interface, no pricing printed. The corporate site shows the lineage: a SenseNova family whose current headline model is an image generator named SenseNova U1 Pro, plus video creation (Seko), an office agent (小浣熊) and digital-human marketing (如影), sold through a token plan. The real strength is the vision lineage: chat that ingests images and video reads native here.

Outside China assume you are an edge case. It expects a mainland account with phone-number and real-name verification, no international payment or English product surface on the pages we read, contracts and data residency in the mainland, and SenseTime's export-control designations, which add a layer for foreign buyers. It loses on text reasoning and cost-performance, where the race in China is run by DeepSeek, Qwen and Kimi; the SenseNova/SenseChat/小浣熊 naming makes it hard to know which checkpoint you are served; and consumer polish trails enterprise delivery. Pick it for image- and video-heavy work inside a Chinese organisation that already buys SenseTime's vision stack — a vertical multimodal vendor, not a general assistant.