Google日前正式发布全新开源大模型Gemma412B，标志着端侧全模态AI迎来突破性进展。该模型颠覆了传统多模态模型依赖视觉和音频外部编码器的复杂链路，创新性地采用“Unified”无编码器架

当图文、视频、图表甚至UI界面都能被统一“理解”并精准匹配，多模态信息检索的边界正在被彻底重构。今日，阿里通义实验室正式开源Qwen3-VL-Embedding与Qwen3-VL-Reranker

Recently, Kuaizhi Wanyi officially released its brand-new open-source model Skywork-R1V3.0, claimi

Recently,昆仑万维Skywork and Nanyang Technological University jointly launched an innovative hierarchi