[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-b115486a-b837-4de1-9dac-d2237723ee85":3},{"id":4,"title":5,"summary":6,"original_url":7,"source_id":8,"tags":9,"published_at":23,"created_at":24,"modified_at":25,"is_published":26,"publish_type":27,"image_url":13,"view_count":28},"b115486a-b837-4de1-9dac-d2237723ee85","宇树 UnifoLM-OminiA-0.3:G1 上跑通\"感知—行动\"端到端大模型","7月20日,宇树科技发布 UnifoLM-OminiA-0.3,把视觉识别、语义理解、精细操作和设备控制四类能力,塞进一个端到端大模型,并在自家的人形机器人 G1 上跑出从感知到行动的完整闭环。\n\n演示视频里,它能把抱枕放上沙发、识别药盒颜色数量并取出指定一盒、调节病床高度并对\"停\"指令即时响应。横跨\"对话—识别—规划—执行\"四个层面,过去要靠视觉、语音、决策、控制四套模型协同,现在一次性打通。\n\n技术上,\"全模态\"终于从 PPT 走到 demo。该模型支持视觉、语音、动作指令联合输入,直接输出机器人运动控制指令,省掉传统 pipeline 的中间表征转换。在康养、家居这种强干扰、跨任务环境里,少一次模态转换就少一次误差累积,抗干扰能力自然水涨船高。\n\n产业上有三层意义:把\"物理 AI\"从论文拽到硬件做闭环;具身智能终于有了可按 demo 评估的基线;这是国内少见的模型+硬件深度耦合的端到端方案,而不是只发 API 让人去接。当然,目前仍是 demo 级表现,长尾场景与对未知物体的泛化,都还需要大规模部署来检验。但 2026 下半年的具身智能赛道,玩家已从\"我能跑\"升级到\"我能稳\",宇树这一步,算是把门槛往上抬了一截。","https:\u002F\u002F36kr.com\u002Fnewsflashes\u002F3903657704277633","5e4fd3d1-9cb4-44a6-bae5-9ffb449c05c1",[10,14,17,20],{"id":11,"name":12,"slug":12,"description":13,"color":13},"a8002d98-9df1-4ab9-94d4-a7625af634c4","china-ai",null,{"id":15,"name":16,"slug":16,"description":13,"color":13},"7e89b5cc-57db-4f37-bc6d-28919a73931c","model-release",{"id":18,"name":19,"slug":19,"description":13,"color":13},"499f4b56-819d-49a3-9609-33e775143b86","multimodal",{"id":21,"name":22,"slug":22,"description":13,"color":13},"b1853a5a-d940-42b7-94f9-0488ee3f2cf7","new-model","2026-07-20T08:01:00Z","2026-07-20T08:03:51.150368Z","2026-07-20T08:03:51.150378Z",true,"agent",8]