{"@context":"https://schema.org","@type":"NewsArticle","generatedAt":"2026-07-23T08:01:28.298Z","headline":"腾讯发布具身 VLM 基座模型 Hy-Embodied-VLM-1.0，A3B 规模性能接近上一代 A32B 模型","description":"腾讯 Robotics X 实验室与福田实验室联合发布第二代具身 VLM 基座模型 Hy-Embodied-VLM-1.0。该模型仅采用 A3B 规模，在 37 个评测任务中综合平均得分 65.6，整体性能接近上一代 A32B 旗舰模型。模型已开源。","url":"https://www.aioga.com/news/cmromti4j04yvbitot8x40gja/","mainEntityOfPage":"https://www.aioga.com/news/cmromti4j04yvbitot8x40gja/","datePublished":"2026-07-17T07:22:18.000Z","dateModified":"2026-07-17T07:22:18.000Z","inLanguage":"zh-CN","publisher":{"@type":"NewsMediaOrganization","name":"Aioga","url":"https://www.aioga.com"},"citation":["https://www.ithome.com/0/978/080.htm","https://aihot.virxact.com/items/cmromti4j04yvbitot8x40gja"],"canonicalUrl":"https://www.aioga.com/news/cmromti4j04yvbitot8x40gja/","directAnswer":{"@type":"Answer","text":"Aioga 编辑摘要：腾讯 Robotics X 实验室与福田实验室联合发布第二代具身 VLM 基座模型 Hy-Embodied-VLM-1.0。 Aioga 将其归入「模型更新」方向，重点关注它对真实使用和行业竞争的影响。","url":"https://www.aioga.com/news/cmromti4j04yvbitot8x40gja/","dateCreated":"2026-07-17T07:22:18.000Z","author":{"@type":"Organization","@id":"https://www.aioga.com/authors/aioga-editorial/#editorial-team","name":"Aioga Editorial Team","url":"https://www.aioga.com/authors/aioga-editorial/"}},"evidence":[{"@type":"CreativeWork","name":"IT之家 source article","url":"https://www.ithome.com/0/978/080.htm","datePublished":"2026-07-17T07:22:18.000Z","provider":{"@type":"Organization","name":"IT之家","url":"https://www.ithome.com/0/978/080.htm"}},{"@type":"CreativeWork","name":"AIHot archive record","url":"https://aihot.virxact.com/items/cmromti4j04yvbitot8x40gja","datePublished":"2026-07-17T07:22:18.000Z","provider":{"@type":"Organization","name":"AIHot","url":"https://aihot.virxact.com/items/cmromti4j04yvbitot8x40gja"}}],"aggregationSource":"IT之家（RSS）","originalPublisher":{"name":"IT之家","url":"https://www.ithome.com/0/978/080.htm"},"article":{"id":"cmromti4j04yvbitot8x40gja","slug":"cmromti4j04yvbitot8x40gja","url":"https://www.aioga.com/news/cmromti4j04yvbitot8x40gja/","title":"腾讯发布具身 VLM 基座模型 Hy-Embodied-VLM-1.0，A3B 规模性能接近上一代 A32B 模型","title_en":"腾讯发布具身 VLM 基座模型 Hy-Embodied-VLM-1.0，A3B 规模整体性能接近上一代 A32B 模型","summary":"腾讯 Robotics X 实验室与福田实验室联合发布第二代具身 VLM 基座模型 Hy-Embodied-VLM-1.0。该模型仅采用 A3B 规模，在 37 个评测任务中综合平均得分 65.6，整体性能接近上一代 A32B 旗舰模型。模型已开源。","source":"IT之家（RSS）","sourceUrl":"https://www.ithome.com/0/978/080.htm","aiHotUrl":"https://aihot.virxact.com/items/cmromti4j04yvbitot8x40gja","publishedAt":"2026-07-17T07:22:18.000Z","category":"模型更新","score":66,"selected":false,"articleBody":["IT之家：https://www.ithome.com/ 7 月 17 日消息，腾讯 Robotics X 实验室、福田实验室联合腾讯混元打造的第二代具身 VLM 基座模型 Hy-Embodied-VLM-1.0 昨日正式发布。","官方表示，在覆盖 37 个评测任务的具身能力评测体系中，Hy-Embodied-VLM-1.0 在物理状态理解、动作 — 变化推理、时序与自适应推理三大维度分别取得 68.6、64.1 和 57.4 分，综合平均得分达到 65.6， 仅采用 A3B 规模，模型整体性能已接近上一代 A32B 旗舰模型 ，并显著优于同等规模的 Qwen3.6-A3B、Cosmos 3-8B、Embodied-R1 通用与具身模型。","据介绍，这一模型从物理空间状态理解、动作-变化理解、时序和自适应推理三个层次构建模型能力维度，增强了场景感知、行动分析规划、导航等面向具身 Agent 的多模态理解能力。","模型能够识别物体、属性、深度与空间关系，理解机器人视角、功能部件、可操作区域及环境可供性，形成与当前任务相关的物理世界状态表征。进一步地，模型可结合人机交互语义、目标对象与任务约束，完成下一步动作选择、目标与轨迹定位，并判断空间前置条件、动作可执行性及其局部物理影响。","面向长时程任务，模型还具备多步规划、视觉语言导航、历史与空间记忆、失败诊断、反事实分析和动态重规划能力，可依据目标、执行进度与环境反馈持续修正决策，实现从场景理解、局部因果推理到长期自适应执行的完整能力覆盖。","https://github.com/Tencent-Hunyuan/HY-Embodied ：https://github.com/Tencent-Hunyuan/HY-Embodied","https://huggingface.co/tencent/Hy-Embodied-VLM-1.0 ：https://huggingface.co/tencent/Hy-Embodied-VLM-1.0"],"articleImages":[{"sourceUrl":"https://img.ithome.com/newsuploadfiles/2026/7/30b708e4-444b-44dc-99c4-447f30f202c2.png?x-bce-process=image/format,f_auto","alt":"","afterParagraph":0,"url":"/media/articles/cmromti4j04yvbitot8x40gja/8974539cf0f9ba5a.webp"},{"sourceUrl":"https://img.ithome.com/newsuploadfiles/2026/7/d3fc610f-9106-4bb7-9cee-9c8b2e58d7b8.png?x-bce-process=image/format,f_auto","alt":"","afterParagraph":4,"url":"/media/articles/cmromti4j04yvbitot8x40gja/cf4da15f79ad04a2.webp"}],"mediaStatus":"ok","articleBodyZh":[],"translationStatus":"","bodyOrigin":"source-page","editorial":{"summary":"Aioga 编辑摘要：腾讯 Robotics X 实验室与福田实验室联合发布第二代具身 VLM 基座模型 Hy-Embodied-VLM-1.0。 Aioga 将其归入「模型更新」方向，重点关注它对真实使用和行业竞争的影响。","background":"背景分析：模型与研究类动态需要结合能力边界、开放方式、成本、可用性和真实任务表现判断，单项指标领先不等于已经形成稳定采用。","viewpoint":"Aioga 判断：这条动态更适合作为行业观察信号，当前信息足以建立线索，但不足以推导长期结论。","implications":"影响分析：对相关团队而言，短期应先核对来源、可用范围和实际成本，再判断是否值得接入或跟进。","nextStep":"后续观察：继续观察官方文档、实际可用性、价格变化、开发者反馈和竞品回应。","evidenceRefs":["title","summary","articleBody"],"confidence":"medium","status":"published","aiGenerated":false,"autoApproved":true,"generatedBy":"rule-safe-fallback","generatedAt":"2026-07-23T08:10:16.840Z","sourceHash":"718dfa0e4d23f0cb","validation":{"passed":true,"mode":"rule-safe-fallback","checks":["schema","length","source-attribution","no-html"]}},"tags":["模型更新","IT之家（RSS）"],"translations":{"zh-CN":{"title":"腾讯发布具身 VLM 基座模型 Hy-Embodied-VLM-1.0，A3B 规模性能接近上一代 A32B 模型","summary":"腾讯 Robotics X 实验室与福田实验室联合发布第二代具身 VLM 基座模型 Hy-Embodied-VLM-1.0。该模型仅采用 A3B 规模，在 37 个评测任务中综合平均得分 65.6，整体性能接近上一代 A32B 旗舰模型。模型已开源。","category":"模型更新","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"腾讯发布具身 VLM 基座模型 Hy-Embodied-VLM-1.0，A3B 规模性能接近上一代 A32B 模型 - Aioga AI资讯","description":"腾讯 Robotics X 实验室与福田实验室联合发布第二代具身 VLM 基座模型 Hy-Embodied-VLM-1.0。该模型仅采用 A3B 规模，在 37 个评测任务中综合平均得分 65.6，整体性能接近上一代 A32B 旗舰模型。模型已开源。","url":"https://www.aioga.com/news/cmromti4j04yvbitot8x40gja/"},"en":{"title":"Tencent releases the embodied VLM base model Hy-Embodied-VLM-1.0, with A3B scale performance close to the previous generation A32B model","summary":"Tencent Robotics X Lab and Futian Lab jointly released the second-generation embodied VLM base model Hy-Embodied-VLM-1.0. This model only uses A3B scale and achieved an average score of 65.6 across 37 evaluation tasks, with overall performance close to the previous generation A32B flagship model. The model has been open-sourced.","category":"Models","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Tencent releases the embodied VLM base model Hy-Embodied-VLM-1.0, with A3B scale performance close to the previous generation A32B model - Aioga AI News","description":"Tencent Robotics X Lab and Futian Lab jointly released the second-generation embodied VLM base model Hy-Embodied-VLM-1.0. This model only uses A3B scale and achieved an average sco...","url":"https://www.aioga.com/en/news/cmromti4j04yvbitot8x40gja/","contentTranslated":true,"sourceHash":"9670295e6a17a0c4","translatedAt":"2026-07-22T19:24:06.791Z"},"ja":{"title":"テンセントは具現化VLM基盤モデル Hy-Embodied-VLM-1.0 を発表、A3B規模の性能は前世代のA32Bモデルに近い","summary":"テンセント Robotics X ラボとフタデンラボは、第2世代の具現化 VLM 基盤モデル Hy-Embodied-VLM-1.0 を共同で発表しました。このモデルは A3B 規模のみを採用しており、37 の評価タスクで総合平均スコア 65.6 を達成し、全体的な性能は前世代の A32B フラッグシップモデルに近いです。モデルはすでにオープンソース化されています。","category":"モデル更新","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"テンセントは具現化VLM基盤モデル Hy-Embodied-VLM-1.0 を発表、A3B規模の性能は前世代のA32Bモデルに近い - Aioga AIニュース","description":"テンセント Robotics X ラボとフタデンラボは、第2世代の具現化 VLM 基盤モデル Hy-Embodied-VLM-1.0 を共同で発表しました。このモデルは A3B 規模のみを採用しており、37 の評価タスクで総合平均スコア 65.6 を達成し、全体的な性能は前世代の A32B フラッグシップモデルに近いです。モデルはすでにオープンソース化されて...","url":"https://www.aioga.com/ja/news/cmromti4j04yvbitot8x40gja/","contentTranslated":true,"sourceHash":"9670295e6a17a0c4","translatedAt":"2026-07-22T19:24:05.865Z"},"ko":{"title":"텐센트, 체현 VLM 기반 모델 Hy-Embodied-VLM-1.0 공개, A3B 규모 성능이 이전 세대 A32B 모델에 근접","summary":"텅쉰 Robotics X 실험실과 푸톈 실험실이 공동으로 제2세대 신체화 VLM 기반 모델 Hy-Embodied-VLM-1.0을 발표했습니다. 이 모델은 A3B 규모만 사용했으며, 37개의 평가 과제에서 종합 평균 점수 65.6점을 기록했으며, 전체 성능은 이전 세대 A32B 플래그십 모델에 근접합니다. 모델은 이미 오픈소스로 공개되었습니다.","category":"모델 업데이트","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"텐센트, 체현 VLM 기반 모델 Hy-Embodied-VLM-1.0 공개, A3B 규모 성능이 이전 세대 A32B 모델에 근접 - Aioga AI 뉴스","description":"텅쉰 Robotics X 실험실과 푸톈 실험실이 공동으로 제2세대 신체화 VLM 기반 모델 Hy-Embodied-VLM-1.0을 발표했습니다. 이 모델은 A3B 규모만 사용했으며, 37개의 평가 과제에서 종합 평균 점수 65.6점을 기록했으며, 전체 성능은 이전 세대 A32B 플래그십 모델에 근접합니다. 모델은 이미 오픈...","url":"https://www.aioga.com/ko/news/cmromti4j04yvbitot8x40gja/","contentTranslated":true,"sourceHash":"9670295e6a17a0c4","translatedAt":"2026-07-22T19:24:58.826Z"},"es":{"title":"Tencent lanza el modelo base VLM corporal Hy-Embodied-VLM-1.0, con un rendimiento de escala A3B cercano al modelo de la generación anterior A32B","summary":"El laboratorio Tencent Robotics X y el laboratorio Futian lanzaron conjuntamente la segunda generación del modelo base VLM incorporado Hy-Embodied-VLM-1.0. Este modelo utiliza únicamente la escala A3B, y obtuvo un puntaje promedio general de 65,6 en 37 tareas de evaluación, con un rendimiento global cercano al del modelo insignia de la generación anterior A32B. El modelo ya es de código abierto.","category":"Modelos","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Tencent lanza el modelo base VLM corporal Hy-Embodied-VLM-1.0, con un rendimiento de escala A3B cercano al modelo de la generación anterior A32B - Aioga Noticias de IA","description":"El laboratorio Tencent Robotics X y el laboratorio Futian lanzaron conjuntamente la segunda generación del modelo base VLM incorporado Hy-Embodied-VLM-1.0. Este modelo utiliza únic...","url":"https://www.aioga.com/es/news/cmromti4j04yvbitot8x40gja/","contentTranslated":true,"sourceHash":"9670295e6a17a0c4","translatedAt":"2026-07-22T19:24:48.748Z"},"fr":{"title":"Tencent publie le modèle de base VLM incarné Hy-Embodied-VLM-1.0, dont la performance à l'échelle A3B est proche de celle du modèle A32B de la génération précédente","summary":"Le laboratoire Tencent Robotics X et le laboratoire Futian ont conjointement publié le deuxième modèle de base VLM incarné Hy-Embodied-VLM-1.0. Ce modèle utilise uniquement l'échelle A3B et obtient un score moyen global de 65,6 sur 37 tâches d'évaluation, avec une performance globale proche de celle du modèle phare de la génération précédente A32B. Le modèle est open source.","category":"Modèles","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Tencent publie le modèle de base VLM incarné Hy-Embodied-VLM-1.0, dont la performance à l'échelle A3B est proche de celle du modèle A32B de la génération précédente - Aioga Actualités IA","description":"Le laboratoire Tencent Robotics X et le laboratoire Futian ont conjointement publié le deuxième modèle de base VLM incarné Hy-Embodied-VLM-1.0. Ce modèle utilise uniquement l'échel...","url":"https://www.aioga.com/fr/news/cmromti4j04yvbitot8x40gja/","contentTranslated":true,"sourceHash":"9670295e6a17a0c4","translatedAt":"2026-07-22T19:25:39.914Z"},"de":{"title":"Tencent veröffentlicht das verkörperte VLM-Basis-Modell Hy-Embodied-VLM-1.0, die Leistung des A3B-Formats nähert sich dem Vorgängermodell A32B","summary":"Tencent Robotics X Lab und Futian Lab haben gemeinsam das zweite Generation des verkörperten VLM-Basismodells Hy-Embodied-VLM-1.0 veröffentlicht. Das Modell verwendet nur das A3B-Maß, erzielte in 37 Bewertungsaufgaben eine durchschnittliche Gesamtnote von 65,6 und seine Gesamtleistung liegt nahe beim vorherigen A32B-Flaggschiffmodell. Das Modell ist Open Source veröffentlicht worden.","category":"模型更新","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Tencent veröffentlicht das verkörperte VLM-Basis-Modell Hy-Embodied-VLM-1.0, die Leistung des A3B-Formats nähert sich dem Vorgängermodell A32B - Aioga KI-News","description":"Tencent Robotics X Lab und Futian Lab haben gemeinsam das zweite Generation des verkörperten VLM-Basismodells Hy-Embodied-VLM-1.0 veröffentlicht. Das Modell verwendet nur das A3B-M...","url":"https://www.aioga.com/de/news/cmromti4j04yvbitot8x40gja/","contentTranslated":true,"sourceHash":"9670295e6a17a0c4","translatedAt":"2026-07-22T19:25:48.669Z"},"pt-BR":{"title":"A Tencent lançou o modelo base VLM corporificado Hy-Embodied-VLM-1.0, com desempenho em escala A3B próximo ao modelo da geração anterior A32B","summary":"O laboratório Tencent Robotics X e o laboratório Futian lançaram conjuntamente a segunda geração do modelo base VLM corporificado Hy-Embodied-VLM-1.0. Este modelo utiliza apenas a escala A3B, obtendo uma pontuação média geral de 65,6 em 37 tarefas de avaliação, com desempenho global próximo ao do modelo premium da geração anterior A32B. O modelo já foi disponibilizado como código aberto.","category":"模型更新","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"A Tencent lançou o modelo base VLM corporificado Hy-Embodied-VLM-1.0, com desempenho em escala A3B próximo ao modelo da geração anterior A32B - Aioga Notícias de IA","description":"O laboratório Tencent Robotics X e o laboratório Futian lançaram conjuntamente a segunda geração do modelo base VLM corporificado Hy-Embodied-VLM-1.0. Este modelo utiliza apenas a...","url":"https://www.aioga.com/pt-BR/news/cmromti4j04yvbitot8x40gja/","contentTranslated":true,"sourceHash":"9670295e6a17a0c4","translatedAt":"2026-07-22T19:26:31.943Z"},"ru":{"title":"Tencent выпустила базовую модель осязаемого VLM Hy-Embodied-VLM-1.0, производительность на масштабе A3B близка к предыдущей модели A32B","summary":"Лаборатория Tencent Robotics X совместно с лабораторией Futian представила второе поколение базовой модели воплощённого VLM Hy-Embodied-VLM-1.0. Эта модель использует только масштаб A3B и набрала средний суммарный балл 65,6 в 37 тестовых заданиях, при этом общая производительность близка к предыдущей флагманской модели A32B. Модель уже открыта.","category":"模型更新","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Tencent выпустила базовую модель осязаемого VLM Hy-Embodied-VLM-1.0, производительность на масштабе A3B близка к предыдущей модели A32B - Aioga Новости ИИ","description":"Лаборатория Tencent Robotics X совместно с лабораторией Futian представила второе поколение базовой модели воплощённого VLM Hy-Embodied-VLM-1.0. Эта модель использует только масшта...","url":"https://www.aioga.com/ru/news/cmromti4j04yvbitot8x40gja/","contentTranslated":true,"sourceHash":"9670295e6a17a0c4","translatedAt":"2026-07-22T19:26:30.581Z"},"ar":{"title":"أصدرت شركة تينسنت نموذج قاعدة VLM المجسَّم Hy-Embodied-VLM-1.0، حيث أن أداء حجم A3B يقترب من نموذج الجيل السابق A32B","summary":"أصدرت مختبرات Tencent Robotics X ومختبرات Foton النموذج الأساسي للجيل الثاني من VLM المجسد Hy-Embodied-VLM-1.0. يستخدم هذا النموذج المقياس A3B فقط، وحقق متوسط ​​درجة 65.6 عبر 37 مهمة تقييمية، وأداءه العام قريب من نموذج الجيل السابق الرائد A32B. النموذج أصبح مفتوح المصدر.","category":"模型更新","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"أصدرت شركة تينسنت نموذج قاعدة VLM المجسَّم Hy-Embodied-VLM-1.0، حيث أن أداء حجم A3B يقترب من نموذج الجيل السابق A32B - Aioga أخبار الذكاء الاصطناعي","description":"أصدرت مختبرات Tencent Robotics X ومختبرات Foton النموذج الأساسي للجيل الثاني من VLM المجسد Hy-Embodied-VLM-1.0. يستخدم هذا النموذج المقياس A3B فقط، وحقق متوسط ​​درجة 65.6 عبر 37 مه...","url":"https://www.aioga.com/ar/news/cmromti4j04yvbitot8x40gja/","contentTranslated":true,"sourceHash":"9670295e6a17a0c4","translatedAt":"2026-07-22T19:27:21.195Z"},"hi":{"title":"टेनसेंट ने जीवित VLM बेस मॉडल Hy-Embodied-VLM-1.0 जारी किया, A3B आकार प्रदर्शन पिछले पीढ़ी के A32B मॉडल के करीब है","summary":"टेनसेंट रोबोटिक्स एक्स प्रयोगशाला और फुतियन प्रयोगशाला ने दूसरी पीढ़ी का एम्बॉडीड VLM बेस मॉडल Hy-Embodied-VLM-1.0 संयुक्त रूप से जारी किया। यह मॉडल केवल A3B पैमाने का उपयोग करता है, 37 मूल्यांकन कार्यों में औसत संयुक्त स्कोर 65.6 हासिल किया, और कुल प्रदर्शन पिछली पीढ़ी के A32B प्रमुख मॉडल के करीब है। मॉडल को ओपन-सोर्स कर दिया गया है।","category":"模型更新","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"टेनसेंट ने जीवित VLM बेस मॉडल Hy-Embodied-VLM-1.0 जारी किया, A3B आकार प्रदर्शन पिछले पीढ़ी के A32B मॉडल के करीब है - Aioga AI समाचार","description":"टेनसेंट रोबोटिक्स एक्स प्रयोगशाला और फुतियन प्रयोगशाला ने दूसरी पीढ़ी का एम्बॉडीड VLM बेस मॉडल Hy-Embodied-VLM-1.0 संयुक्त रूप से जारी किया। यह मॉडल केवल A3B पैमाने का उपयोग करता ह...","url":"https://www.aioga.com/hi/news/cmromti4j04yvbitot8x40gja/","contentTranslated":true,"sourceHash":"9670295e6a17a0c4","translatedAt":"2026-07-22T19:27:21.840Z"},"it":{"title":"Tencent ha rilasciato il modello base VLM incarnato Hy-Embodied-VLM-1.0, con prestazioni su scala A3B prossime al modello A32B della generazione precedente","summary":"Il laboratorio Tencent Robotics X e il laboratorio Futian hanno rilasciato congiuntamente il secondo modello di base VLM incarnato Hy-Embodied-VLM-1.0. Questo modello utilizza solo la scala A3B e ottiene un punteggio medio complessivo di 65,6 su 37 compiti di valutazione, con prestazioni complessive vicine al modello di punta della generazione precedente A32B. Il modello è stato reso open source.","category":"模型更新","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Tencent ha rilasciato il modello base VLM incarnato Hy-Embodied-VLM-1.0, con prestazioni su scala A3B prossime al modello A32B della generazione precedente - Aioga Notizie IA","description":"Il laboratorio Tencent Robotics X e il laboratorio Futian hanno rilasciato congiuntamente il secondo modello di base VLM incarnato Hy-Embodied-VLM-1.0. Questo modello utilizza solo...","url":"https://www.aioga.com/it/news/cmromti4j04yvbitot8x40gja/","contentTranslated":true,"sourceHash":"9670295e6a17a0c4","translatedAt":"2026-07-22T19:28:03.635Z"},"nl":{"title":"Tencent heeft het belichaamde VLM-basismodel Hy-Embodied-VLM-1.0 uitgebracht, met A3B-schaalprestaties die dicht bij het vorige generatie A32B-model liggen","summary":"Tencent Robotics X Lab en Futian Lab hebben gezamenlijk het tweede generatie belichaamde VLM-basismodel Hy-Embodied-VLM-1.0 uitgebracht. Dit model gebruikt slechts A3B-schaal en behaalde een gemiddelde score van 65,6 over 37 evaluatietaken, met een algehele prestatie die dicht bij het vorige generatie A32B vlaggenschipmodel ligt. Het model is open-source gemaakt.","category":"模型更新","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Tencent heeft het belichaamde VLM-basismodel Hy-Embodied-VLM-1.0 uitgebracht, met A3B-schaalprestaties die dicht bij het vorige generatie A32B-model liggen - Aioga AI-nieuws","description":"Tencent Robotics X Lab en Futian Lab hebben gezamenlijk het tweede generatie belichaamde VLM-basismodel Hy-Embodied-VLM-1.0 uitgebracht. Dit model gebruikt slechts A3B-schaal en be...","url":"https://www.aioga.com/nl/news/cmromti4j04yvbitot8x40gja/","contentTranslated":true,"sourceHash":"9670295e6a17a0c4","translatedAt":"2026-07-22T19:28:02.714Z"},"tr":{"title":"Tencent, vücut odaklı VLM temel modeli Hy-Embodied-VLM-1.0'i yayınladı, A3B ölçeğindeki performans bir önceki nesil A32B modeliyle yakın.","summary":"Tencent Robotics X laboratuvarı ve Foton laboratuvarı, ikinci nesil vücut tabanlı VLM modeli Hy-Embodied-VLM-1.0'i ortak olarak yayınladı. Bu model yalnızca A3B ölçeğini kullanıyor ve 37 değerlendirme görevinde ortalama toplam puan 65,6 ile önceki nesil A32B amiral gemisi modeline yakın bir genel performans sergiliyor. Model açık kaynak olarak yayınlandı.","category":"模型更新","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Tencent, vücut odaklı VLM temel modeli Hy-Embodied-VLM-1.0'i yayınladı, A3B ölçeğindeki performans bir önceki nesil A32B modeliyle yakın. - Aioga AI Haberleri","description":"Tencent Robotics X laboratuvarı ve Foton laboratuvarı, ikinci nesil vücut tabanlı VLM modeli Hy-Embodied-VLM-1.0'i ortak olarak yayınladı. Bu model yalnızca A3B ölçeğini kullanıyor...","url":"https://www.aioga.com/tr/news/cmromti4j04yvbitot8x40gja/","contentTranslated":true,"sourceHash":"9670295e6a17a0c4","translatedAt":"2026-07-22T19:28:52.097Z"},"vi":{"title":"Tencent phát hành mô hình cơ sở VLM hiện thân Hy-Embodied-VLM-1.0, hiệu năng quy mô A3B gần bằng mô hình A32B thế hệ trước","summary":"Phòng thí nghiệm Tencent Robotics X và phòng thí nghiệm Futian đã cùng phát hành mô hình cơ sở VLM thế hệ thứ hai có thể thực hiện tác vụ thể hiện Hy-Embodied-VLM-1.0. Mô hình này chỉ sử dụng quy mô A3B, đạt điểm trung bình tổng hợp 65,6 trong 37 nhiệm vụ đánh giá, hiệu suất tổng thể gần bằng mô hình chủ lực thế hệ trước A32B. Mô hình đã được mở nguồn.","category":"模型更新","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Tencent phát hành mô hình cơ sở VLM hiện thân Hy-Embodied-VLM-1.0, hiệu năng quy mô A3B gần bằng mô hình A32B thế hệ trước - Tin tức AI Aioga","description":"Phòng thí nghiệm Tencent Robotics X và phòng thí nghiệm Futian đã cùng phát hành mô hình cơ sở VLM thế hệ thứ hai có thể thực hiện tác vụ thể hiện Hy-Embodied-VLM-1.0. Mô hình này...","url":"https://www.aioga.com/vi/news/cmromti4j04yvbitot8x40gja/","contentTranslated":true,"sourceHash":"9670295e6a17a0c4","translatedAt":"2026-07-22T19:28:48.667Z"},"id":{"title":"Tencent merilis model dasar VLM berbadan Hy-Embodied-VLM-1.0, dengan kinerja skala A3B mendekati model A32B generasi sebelumnya","summary":"Laboratorium Robotics X Tencent dan Laboratorium Futian secara bersama-sama merilis model basis VLM berwujud generasi kedua, Hy-Embodied-VLM-1.0. Model ini hanya menggunakan skala A3B, dengan skor rata-rata gabungan 65,6 pada 37 tugas evaluasi, dan kinerja keseluruhan mendekati model andalan generasi sebelumnya A32B. Model ini telah dibuka sumbernya.","category":"模型更新","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Tencent merilis model dasar VLM berbadan Hy-Embodied-VLM-1.0, dengan kinerja skala A3B mendekati model A32B generasi sebelumnya - Berita AI Aioga","description":"Laboratorium Robotics X Tencent dan Laboratorium Futian secara bersama-sama merilis model basis VLM berwujud generasi kedua, Hy-Embodied-VLM-1.0. Model ini hanya menggunakan skala...","url":"https://www.aioga.com/id/news/cmromti4j04yvbitot8x40gja/","contentTranslated":true,"sourceHash":"9670295e6a17a0c4","translatedAt":"2026-07-22T19:29:36.522Z"},"th":{"title":"เทนเซนต์เปิดตัวโมเดลฐาน VLM แบบมีกายภาพ Hy-Embodied-VLM-1.0 ขนาด A3B มีประสิทธิภาพใกล้เคียงกับโมเดล A32B รุ่นก่อนหน้า","summary":"ห้องทดลอง Tencent Robotics X ร่วมกับห้องทดลอง Foton เปิดตัวรุ่นที่สองของโมเดลฐาน VLM แบบมีร่างกาย Hy-Embodied-VLM-1.0 โมเดลนี้ใช้ขนาดเพียง A3B โดยทำคะแนนเฉลี่ยรวม 65.6 ใน 37 งานทดสอบ ประสิทธิภาพโดยรวมใกล้เคียงกับโมเดลเรือธงรุ่นก่อนหน้า A32B โมเดลได้เปิดซอร์สแล้ว","category":"模型更新","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"เทนเซนต์เปิดตัวโมเดลฐาน VLM แบบมีกายภาพ Hy-Embodied-VLM-1.0 ขนาด A3B มีประสิทธิภาพใกล้เคียงกับโมเดล A32B รุ่นก่อนหน้า - ข่าว AI Aioga","description":"ห้องทดลอง Tencent Robotics X ร่วมกับห้องทดลอง Foton เปิดตัวรุ่นที่สองของโมเดลฐาน VLM แบบมีร่างกาย Hy-Embodied-VLM-1.0 โมเดลนี้ใช้ขนาดเพียง A3B โดยทำคะแนนเฉลี่ยรวม 65.6 ใน 37 งานทดส...","url":"https://www.aioga.com/th/news/cmromti4j04yvbitot8x40gja/","contentTranslated":true,"sourceHash":"9670295e6a17a0c4","translatedAt":"2026-07-22T19:29:43.266Z"},"pl":{"title":"Tencent wypuścił model bazowy Hy-Embodied-VLM-1.0 VLM z ciałem, wydajność w skali A3B zbliżona do poprzedniej generacji modelu A32B","summary":"Laboratorium Tencent Robotics X we współpracy z laboratorium Futian ogłosiło drugą generację uosobionego modelu bazowego VLM Hy-Embodied-VLM-1.0. Model ten wykorzystuje jedynie skalę A3B, osiągając średni wynik 65,6 w 37 zadaniach testowych, a jego ogólna wydajność jest zbliżona do poprzedniej flagowej generacji A32B. Model został udostępniony jako open source.","category":"模型更新","source":"IT之家（RSS）","aggregationSource":"IT之家（RSS）","pageTitle":"Tencent wypuścił model bazowy Hy-Embodied-VLM-1.0 VLM z ciałem, wydajność w skali A3B zbliżona do poprzedniej generacji modelu A32B - Aioga Wiadomości AI","description":"Laboratorium Tencent Robotics X we współpracy z laboratorium Futian ogłosiło drugą generację uosobionego modelu bazowego VLM Hy-Embodied-VLM-1.0. Model ten wykorzystuje jedynie ska...","url":"https://www.aioga.com/pl/news/cmromti4j04yvbitot8x40gja/","contentTranslated":true,"sourceHash":"9670295e6a17a0c4","translatedAt":"2026-07-22T19:30:33.655Z"}}}}