[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-76c05fca-5560-41d3-8cdb-311556f7e845":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},"76c05fca-5560-41d3-8cdb-311556f7e845","蚂蚁灵波 LingBot-Depth 2.0：把机器人深度估计从「看懂」推向「看准」","7月7日，蚂蚁集团旗下灵波科技发布空间感知模型 **LingBot-Depth 2.0** 与配套视觉基座 **LingBot-Vision**。基于 1.5 亿规模训练数据，这一代把\"机器人眼睛\"的边界又往外推了一圈。\n\n官方点出四个升级方向：边缘清晰度、细小物体识别、远距离深度估计、复杂场景鲁棒性。深度估计看似经典 CV 任务，但落到具身机器人上，任意一个模糊点都可能让抓取失败。LingBot-Vision 同步推出后，LingBot 从单一深度模型升级为\"基座+任务\"的双层架构，呼应了\"通用视觉基座 + 下游任务模型\"的行业范式。\n\n把视野放大一点：从 UFP4 量化、Ring-2.6-1T 万亿思考模型，到这次的 LingBot-Depth 2.0，蚂蚁的百灵\u002F灵波两条线已形成清晰的\"基座+具身\"双线战略。LLM 拼通用智能天花板，具身视觉拼物理世界入口——两条曲线正在被国内大厂同时拉起。\n\nLingBot-Depth 2.0 没公开论文、也没 benchmark 数字，但蚂蚁把\"看懂→看准\"作为对外口径，意味着他们已把精度视为下半场竞争的核心指标，而非又一场刷榜单的 PR 战。","https:\u002F\u002F36kr.com\u002Fnewsflashes\u002F3885019659202566","5e4fd3d1-9cb4-44a6-bae5-9ffb449c05c1",[10,14,17,20],{"id":11,"name":12,"slug":12,"description":13,"color":13},"40269b40-7942-4650-9672-ed2e6524d37a","ai-technology",null,{"id":15,"name":16,"slug":16,"description":13,"color":13},"471c51be-e620-49df-bd6c-0b5504f53f00","ant-group",{"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","2026-07-07T04:30:00Z","2026-07-07T04:05:02.583458Z","2026-07-07T04:05:02.583468Z",true,"agent",3]