PulseAugur
EN
LIVE 21:31:56

NVIDIA launches Cosmos 3 for physical AI, bridging perception and action

NVIDIA has unveiled Cosmos 3, a new foundational model designed for physical AI applications. This model aims to bridge the gap between perception and action, enabling AI systems to understand and interact with the physical world. Cosmos 3 is touted for its high performance in image and video generation and is expected to be crucial for advancements in robotics and autonomous driving. AI

IMPACT Enables AI systems to better understand and interact with the physical world, accelerating robotics and autonomous driving development.

RANK_REASON Frontier-lab model release with system card.

Read on MarkTechPost →

AI-generated summary · Google Gemini · from 6 sources. How we write summaries →

NVIDIA launches Cosmos 3 for physical AI, bridging perception and action

COVERAGE [6]

  1. MarkTechPost TIER_1 English(EN) · Asif Razzaq ·

    NVIDIA Releases Cosmos 3: A Two-Tower Mixture-of-Transformers Foundation Model Unifying Physical Reasoning, World Generation, and Action Generation

    <p>NVIDIA released Cosmos 3, open omnimodal world models pairing an autoregressive VLM reasoner with a diffusion generator for physical AI.</p> <p>The post <a href="https://www.marktechpost.com/2026/06/03/nvidia-releases-cosmos-3-a-two-tower-mixture-of-transformers-foundation-mod…

  2. Mastodon — fosstodon.org TIER_1 日本語(JA) · [email protected] ·

    📝 NVIDIA's 'Perception to Action' with 'Cosmos 3' - Physical AI Begins to Understand the Physical World. NVIDIA's new foundational model 'Cosmos 3' achieves top performance in image and video generation. The era of Physical AI, serving as the 'eyes and brain' for robots and autonomous driving, is truly beginning. 🔗 https://techscope365.com

    📝 NVIDIAが「Cosmos 3」で実現する「知覚から行動へ」——フィジカルAIが物理世界を理解し始めた NVIDIAの新基盤モデル「Cosmos 3」は、画像・動画生成の最高性能を実現。ロボットや自動運転の「目と脳」となるフィジカルAIの時代が本格化します。 🔗 https:// techscope365.com/803/ # フィジカルAI # NVIDIACosmos # 画像生成 # AI # テクノロジー

  3. Mastodon — fosstodon.org TIER_1 English(EN) · [email protected] ·

    https:// winbuzzer.com/2026/06/01/nvidi a-launches-cosmos-3-with-openmdw-for-physical-ai-xcxwbn/ NVIDIA has launched Cosmos 3 as a physical-AI model that combin

    https:// winbuzzer.com/2026/06/01/nvidi a-launches-cosmos-3-with-openmdw-for-physical-ai-xcxwbn/ NVIDIA has launched Cosmos 3 as a physical-AI model that combines scene reasoning, multimodal generation and action output, tying the release to a new OpenMDW licensing framework. # A…

  4. Mastodon — mastodon.social TIER_1 Italiano(IT) · tomshw ·

    🤖 Nvidia focuses on physical AI: from world models to open humanoid robots, AI leaves screens and learns to move in the real world. # Nvidia # AI 🔗 https://w

    🤖 Nvidia punta sulla physical AI: dai world model ai robot umanoidi open, l’AI esce dagli schermi e impara a muoversi nel mondo reale. # Nvidia # AI 🔗 https://www. tomshw.it/business/nvidia-gioc a-tutto-sulla-physical-ai-dal-world-model-al-robot-umanoide-open-2026-06-02

  5. Mastodon — mastodon.social TIER_1 Polski(PL) · aisight ·

    Nvidia Cosmos is an open-source world model that teaches AI systems the laws of physics. Thanks to it, robots and autonomous cars will be able to train in perfectly replicated environments.

    Nvidia Cosmos to otwartoźródłowy model świata, który uczy systemy AI zasad fizyki. Dzięki niemu roboty i autonomiczne auta będą mogły trenować w idealnie odwzorowanych symulacjach przed wyjazdem na ulice. # si # ai # sztucznainteligencja # wiadomości # informacje # technologia ht…

  6. Mastodon — mastodon.social TIER_1 日本語(JA) · [email protected] ·

    NVIDIA's World Model "Cosmos 3" for Physical AI, Including 3 Types for Edge, etc. https://www.watch.impress.co.jp/docs/news/113492.html # watch_impress # tech # AI

    NVIDIA、フィジカルAIのための世界モデル「Cosmos 3」 エッジ用など3種類 https://www. watch.impress.co.jp/docs/news/ 2113492.html # watch_impress # テック # AI