# Physics Published articles for Physics. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## Decoding cosmic signals with deep learning and Keras DevFeed: [Decoding cosmic signals with deep learning and Keras]() Original publisher: [Read original article]() Author: Yufeng Guo; Jonas Glombitza, PhD Published: 2026-09-12T11:04:33.891311Z Content type: article Language: en Sources: [Google Developers Blog]() Topics: [Deep learning](), [Keras](), [data]() Tags: [analysis](), [blog-post](), [data](), [deep-learning](), [keras](), [particle-physics](), [physics](), [research]() ### AI overview The article explains how deep learning and Keras can help analyze the enormous, complex datasets produced by astroparticle-physics observatories. These methods may improve instrument sensitivity, reveal hidden patterns, and identify anomalies in signals from cosmic messengers such as photons, neutrinos, and cosmic rays. ### Source excerpt Astroparticle physics sits at the exciting intersection of astrophysics and particle physics and stu... ## Prison-Break Comedy Sandbox Bad Inmate Coming Soon To Quest & PlayStation VR2 DevFeed: [Prison-Break Comedy Sandbox Bad Inmate Coming Soon To Quest & PlayStation VR2]() Original publisher: [Read original article]() Author: James Tocchio Published: 2026-08-31T15:43:03Z Content type: news Language: en Sources: [UploadVR]() Topics: [simulator]() Tags: [games](), [physics](), [release](), [sandbox](), [simulator](), [vr-gaming]() ### AI overview HyperVR Games announced that Bad Inmate, a physics-driven comedic prison-break sandbox game, will launch for Meta Quest and PlayStation VR2 on October 15. A SteamVR release is planned, but its date has not been announced. ### Source excerpt The physics-driven comedic prison-break sandbox game is coming to Quest and PlayStation VR2 in October. ## AI helps design new materials that work in the real world DevFeed: [AI helps design new materials that work in the real world]() Original publisher: [Read original article]() Author: Zach Winn | MIT News Published: 2026-08-26T09:00:00Z Content type: news Language: en Sources: [MIT AI News]() Topics: [Artificial Intelligence](), [Framework](), [Simulation and Design](), [Crystal]() Tags: [ai](), [artificial-intelligence](), [bowen-yu](), [chemical-engineering](), [chemistry](), [computer-chips](), [computer-science-and-technology](), [crystal](), [crysvcd](), [data-centers](), [department-of-energy-doe](), [diffusion-models](), [dmse](), [hao-tang](), [heather-kulik](), [ju-li](), [machine-learning](), [materials-design](), [materials-discovery](), [materials-science-and-engineering](), [mingda-li](), [mouyang-cheng](), [national-science-foundation-nsf](), [nuclear-science-and-engineering](), [paper](), [physics](), [research](), [school-of-engineering](), [school-of-science](), [semiconductors](), [weiliang-luo](), [weiwei-xie](), [yongqiang-cheng]() ### AI overview MIT researchers developed CrysVCD, a framework that applies chemistry-based valence constraints before material generation to improve the stability of generated designs. In tests, it achieved high lattice-dynamics stability in nearly 70 percent of computational material generations and supported targeting properties such as high thermal conductivity and high dielectric constant. ### Source excerpt The "CrysVCD" tool developed at MIT could cut the huge amounts of time and money spent on screening out chemically unstable designs. ## Talking with Synopsys about the Physics of Chip Design at DAC 2026 DevFeed: [Talking with Synopsys about the Physics of Chip Design at DAC 2026]() Original publisher: [Read original article]() Author: George Cozma Published: 2026-08-11T16:52:16Z Content type: article Language: en Sources: [Chips and Cheese]() Topics: [Chip design](), [3D](), [Simulation and Design]() Tags: [3d](), [audio](), [automotive](), [battery](), [chip-design](), [communications](), [design](), [desktop](), [heat](), [interview](), [low-power](), [physics](), [techniques]() ### AI overview An edited audio interview with Synopsys executive Ravi Subramanian examines the physics involved in chip design and electronic design automation tools. The discussion covers larger chips, 2.5D and 3D multi-die integration, thermal management, low-power mobile systems, and automotive operating environments. ### Source excerpt Hello you fine Internet folks, ## GeoPT helps AI models simulate how objects respond to physical forces DevFeed: [GeoPT helps AI models simulate how objects respond to physical forces]() Original publisher: [Read original article]() Author: Alex Shipps | MIT CSAIL Published: 2026-08-10T19:25:00Z Content type: news Language: en Sources: [MIT AI News]() Topics: [AI Models](), [Simulation](), [AI Research](), [Computer Science and Artificial Intelligence Laboratory (CSAIL)](), [Robotics]() Tags: [3-d-imaging](), [ai](), [ai-models](), [artificial-intelligence](), [computational-fluid-dynamics-cfd](), [computer-graphics](), [computer-modeling](), [computer-science-and-artificial-intelligence-laboratory-csail](), [computer-science-and-technology](), [crash-simulation](), [design](), [electrical-engineering-and-computer-science-eecs](), [geometric-pre-training](), [geopt](), [haixu-wu](), [human-computer-interaction](), [kaiming-he](), [manufacturing](), [minghao-guo](), [mit-csail](), [mit-eecs](), [mit-schwarzman-college-of-computing](), [neural-physics-simulation](), [paper](), [physics](), [physics-aware-ai](), [physics-foundation-models](), [research](), [robotics](), [school-of-engineering](), [self-supervised-learning](), [simulation](), [surrogate-modeling](), [synthetic-dynamics](), [transformer-based-simulators](), [wojciech-matusik]() ### AI overview Researchers at MIT CSAIL and Tsinghua University developed GeoPT, a pre-training approach that uses 3D simulations of mechanical interactions to help AI models learn physics more efficiently. The article reports that models using the approach reached peak performance twice as fast and trained on up to 60 percent less data than leading models. ### Source excerpt "GeoPT" helps AI models understand the basics of physics so they can simulate how objects respond to things like wind and water more efficiently and accurately. ## Relative velocity and closing speed DevFeed: [Relative velocity and closing speed]() Original publisher: [Read original article]() Author: Eli Bendersky Published: 2026-08-04T03:01:00Z Content type: tutorial Language: en Sources: [Eli Bendersky]() Topics: [Simulation](), [Game engine](), [3D]() Tags: [3d](), [game](), [math](), [misc](), [physics]() ### AI overview A tutorial on closing speed, defined as the normal component of the relative velocity between two objects. It explains vector projection, the role of the connecting line between the objects, and how the sign indicates whether they are approaching or separating. ### Source excerpt In Physics simulations or game engines it's sometimes useful to determine the speed with which two objects are approaching each other. This post will discuss the concept of closing speed, which is the normal component of the relative velocity of two objects. Relative velocity and its components Suppose we ... ## The search for quantum advantage in differential equations DevFeed: [The search for quantum advantage in differential equations]() Original publisher: [Read original article]() Author: Robert Davis Published: 2026-08-03T13:00:00Z Content type: article Language: en Sources: [IBM Research]() Topics: [Quantum Computing](), [Algorithms, Complexity](), [Complex Systems]() Tags: [algorithms](), [complex-systems](), [ibm](), [mathematical-sciences](), [physics](), [q-a](), [quantum](), [quantum-algorithms](), [research](), [systems]() ### AI overview IBM researcher Hari Krovi discusses quantum algorithms for solving certain differential equations and their potential to scale beyond classical methods in selected applications. ### Source excerpt New quantum algorithms could unlock faster ways to model the complex systems behind circuits, fluids, finance, and more. ## Geometry Nodes Physics DevFeed: [Geometry Nodes Physics]() Original publisher: [Read original article]() Author: Jacques Lucke Published: 2026-07-30T14:52:49Z Content type: article Language: en Sources: [Blender]() Topics: [Simulation](), [Framework](), [systems]() Tags: [blender](), [collection](), [experimental](), [features](), [general-development](), [geometry-nodes](), [gravity](), [lts](), [mesh](), [node](), [physics](), [procedural](), [systems](), [vectors](), [workflow]() ### AI overview This article describes Blender 5.2 LTS's new experimental hair and cloth dynamics system built with Geometry Nodes. It explains the declarative XPBD simulation framework, cloth and hair workflows, geometry bundles, and customizable effectors such as colliders, custom forces, and custom behavior closures. ### Source excerpt Geometry Nodes Physics in Blender 5.2 LTS and beyond. ## Developing Healthcare Robotics with GPU-Native Medical Physics Simulation DevFeed: [Developing Healthcare Robotics with GPU-Native Medical Physics Simulation]() Original publisher: [Read original article]() Author: Michelle Horton Published: 2026-07-28T20:49:21Z Content type: article Language: en Sources: [NVIDIA Developer](), [NVIDIA Technical Blog]() Topics: [Isaac for Healthcare](), [Robotics](), [Reinforcement learning](), [GPU](), [datasets](), [Medical imaging](), [Cosmos](), [Open Source](), [real-time](), [Nvidia]() Tags: [agentic-ai-generative-ai](), [data](), [datasets](), [featured](), [gpu](), [healthcare](), [isaac](), [isaac-for-healthcare](), [isaac-sim](), [medical-imaging](), [nvidia](), [open-source](), [physical-ai](), [physics](), [real-time](), [reinforcement-learning](), [robotics](), [simulation-modeling-design](), [warp](), [world-model]() ### AI overview The article presents NVIDIA's open source, GPU-accelerated Medical Physics Simulation framework for healthcare robotics. It addresses limited medical robotics data, poor generalization, and slow development by enabling anatomical digital twins, device-anatomy and medical imaging simulation, and GPU-scale reinforcement learning within Isaac for Healthcare, Isaac Sim, and Isaac Lab. ### Source excerpt Unlike autonomous driving or industrial robotics, healthcare robotics can't rely on internet-scale data collection or unlimited real-world experimentation.... ## Advancing Semiconductor Innovation Across Materials Engineering and Manufacturing DevFeed: [Advancing Semiconductor Innovation Across Materials Engineering and Manufacturing]() Original publisher: [Read original article]() Author: Tanya Lenz Published: 2026-07-27T00:45:00Z Content type: article Language: en Sources: [NVIDIA Developer](), [NVIDIA Technical Blog]() Topics: [Artificial Intelligence](), [CUDA](), [GPU](), [Optimization](), [Nvidia](), [Simulation and Design](), [Hardware]() Tags: [ai](), [computational-chemistry-materials-science](), [compute](), [cuda](), [cuda-x](), [data-center-cloud](), [engineering](), [featured](), [gpu](), [hardware](), [industrial-digitalization-digital-twin](), [manufacturing](), [nvidia](), [optimization](), [physics](), [production](), [simulation-modeling-design]() ### AI overview Applied Materials and NVIDIA are presented as combining materials engineering, semiconductor manufacturing, CUDA-X libraries, GPU-accelerated simulation, physics-based modeling, and AI-driven digital twins in an end-to-end digital development model. The approach spans atomic-scale materials discovery, process development, and factory optimization, with Ginestra used to connect material defects and properties to predicted device performance. ### Source excerpt As AI workloads increase, explosive compute demand is pushing the semiconductor industry to meet unprecedented performance targets. Even small delays can have... ## MIT projects selected for funding under US Department of Energy's Genesis Mission DevFeed: [MIT projects selected for funding under US Department of Energy's Genesis Mission]() Original publisher: [Read original article]() Author: Office of the Vice President for Research Published: 2026-07-23T12:00:00Z Content type: news Language: en Sources: [MIT AI News]() Topics: [Artificial Intelligence](), [Supercomputing](), [systems]() Tags: [administration](), [ai](), [artificial-intelligence](), [center-for-computational-science-and-engineering](), [chemical-engineering](), [department-of-energy](), [department-of-energy-doe](), [doe](), [eaps](), [engineering](), [faculty](), [funding](), [genesis](), [genesis-mission](), [industry](), [initiative](), [laboratory-for-nuclear-science](), [manufacturing](), [mathematics](), [mechanical-engineering](), [mission](), [national-security](), [nuclear-science-and-engineering](), [physics](), [plasma-science-and-fusion-center](), [projects](), [quantum](), [quantum-computing](), [quantum-systems](), [research](), [research-laboratory-of-electronics](), [school-of-engineering](), [school-of-science](), [science](), [supercomputing](), [us]() ### AI overview MIT researchers will contribute to 15 collaborative projects selected for funding under Phase I of the U.S. Department of Energy's Genesis Mission. The projects apply AI, supercomputing, quantum systems, and scientific instruments to research areas including energy, materials, fusion, and national security. Funding remains pending completion of award negotiations. ### Source excerpt Initial research projects advance national priorities across natural resources, manufacturing, nuclear physics, and more. ## How to Evaluate General-Purpose Robot Policies for Real-World Deployment DevFeed: [How to Evaluate General-Purpose Robot Policies for Real-World Deployment]() Original publisher: [Read original article]() Author: Brad Nemire Published: 2026-07-12T01:08:17Z Content type: article Language: en Sources: [NVIDIA Developer](), [NVIDIA Technical Blog]() Topics: [Robotics](), [Benchmark](), [foundation-models](), [Simulation and Design](), [Deployment]() Tags: [benchmark](), [benchmarks](), [deployment](), [evaluation](), [featured](), [foundation-models](), [nvidia-research](), [physics](), [robotics](), [simulation-modeling-design](), [testing]() ### AI overview This developer article examines the challenge of rigorously evaluating general-purpose robot policies for real-world deployment. It discusses simulation as a scalable proxy for expensive real-world testing and identifies limitations in current benchmarks, including shared visual sources between training and evaluation, costly Real2sim reconstruction, static task sets, performance saturation, limited failure diagnostics, and uncertainty in success-rate estimates. ### Source excerpt Robotics foundation models have made remarkable progress. Today's best systems can follow natural language instructions to pick, place, sort, and manipulate a... ## Amazon and University of Michigan give robots a sense of touch DevFeed: [Amazon and University of Michigan give robots a sense of touch]() Original publisher: [Read original article]() Author: Mani Nambi; Nima Fazeli Published: 2026-07-10T17:13:31Z Content type: article Language: en Sources: [Amazon Science homepage]() Topics: [contact-rich manipulation]() Tags: [amazon](), [amazon-robotics-research](), [contact-rich-manipulation](), [dataset-development](), [dexterous-robot-manipulation](), [gelsight-mini-sensor](), [hydroelastic-contact-model](), [hydroshear-simulator](), [physics](), [reinforcement-learning](), [reinforcement-learning-robotics](), [robot-grasping-simulation](), [robot-sense-of-touch](), [robotic-manipulation](), [robotics](), [robots](), [sim-to-real-transfer-robotics](), [training]() ### AI overview HydroShear is a physics-based tactile-force simulator that trains robots for dexterous, contact-rich manipulation in simulation and transfers the resulting policies to real-world tasks. ### Source excerpt HydroShear, a new physics-based simulator, teaches robots how to use their sense of touch to perform complex manipulation tasks, in a way that transfers seamlessly to the real world. ## Jesse Thaler named director of the Laboratory for Nuclear Science DevFeed: [Jesse Thaler named director of the Laboratory for Nuclear Science]() Original publisher: [Read original article]() Author: Julia C. Keller | School of Science Published: 2026-07-07T14:45:00Z Content type: news Language: en Sources: [MIT AI News]() Topics: [Artificial Intelligence](), [AI Research](), [AI, ML & Data Engineering]() Tags: [ai](), [ai-in-physics](), [artificial-intelligence](), [bolek-wyslouch](), [data](), [department-of-energy-doe](), [faculty](), [idss](), [jesse-thaler](), [laboratory-for-nuclear-science](), [leadership](), [mit-center-for-theoretical-physics-ctp](), [mit-ctp-li](), [mit-faculty-appointments](), [mit-iaifi](), [mit-idss](), [mit-leadership](), [mit-lns](), [national-science-foundation-nsf](), [nergis-mavalvala](), [particle-physics](), [physics](), [quantum](), [research](), [school-of-science](), [science](), [tracey-slatyer]() ### AI overview Professor Jesse Thaler has been named director of MIT's Laboratory for Nuclear Science, effective Aug. 1. He succeeds Bolek Wyslouch and will lead the laboratory while continuing research combining particle physics and machine learning. ### Source excerpt The professor of physics and inaugural director of the NSF AI Institute for Artificial Intelligence and Fundamental Interactions will lead LNS and continue his research in particle physics. ## Build Games In Java: Sprites, Box2D Physics And Low-Latency Sound DevFeed: [Build Games In Java: Sprites, Box2D Physics And Low-Latency Sound]() Original publisher: [Read original article]() Author: Shai Almog Published: 2026-06-14T00:00:00Z Content type: article Language: en Sources: [CodeName One]() Topics: [Game Development](), [Java](), [cross-platform](), [Android](), [iOS](), [Simulation]() Tags: [android](), [cross-platform](), [game-development](), [games](), [java](), [low-latency](), [physics](), [windows]() ### AI overview Codename One introduces a com.codename1.gaming package for Java game development, adding a game loop, sprites, pollable input, low-latency sound, and Box2D-based rigid-body physics. The APIs use native compilation to target iOS, Android, and Windows without a JVM at runtime. ### Source excerpt The new com.codename1.gaming package adds a game loop, sprites, pollable input, a low-latency sound pool, and rigid-body physics powered by a bundled Box2D engine, all running unchanged on every platform including iOS. ## AI Now Summit 2026 DevFeed: [AI Now Summit 2026]() Original publisher: [Read original article]() Published: 2026-05-28T12:00:20Z Content type: news Language: en Sources: [Mistral AI Blog]() Topics: [Artificial Intelligence](), [Security](), [datacenter](), [multimodal](), [data](), [Robotics](), [Flight]() Tags: [agentic-ai](), [ai](), [automotive](), [data-center](), [design](), [industry](), [innovation](), [multimodal](), [partner](), [partnership](), [physics](), [production](), [robotics](), [security]() ### AI overview Mistral presents an AI stack for industrial engineering, combining physics models, engineering expertise, and robotics to support design, simulation, and production. Partnerships with Airbus, BMW, and ASML target aircraft operations, crash simulation, and semiconductor engineering, with an emphasis on proprietary-data security. The company also describes Vibe as a unified agent for long-running coding and research tasks and announces a new 10 MW data center in France. ### Source excerpt Innovations for global enterprises solving the world's hardest problems. ## Introducing physics AI at Mistral: the foundation for engineering acceleration. DevFeed: [Introducing physics AI at Mistral: the foundation for engineering acceleration.]() Original publisher: [Read original article]() Published: 2026-05-27T12:00:55Z Content type: article Language: en Sources: [Mistral AI Blog]() Topics: [Artificial Intelligence](), [AI Models](), [Simulation and Design](), [Finite Element Method (FEM)](), [Frontier AI](), [Hardware](), [Deployment]() Tags: [ai](), [ai-data-centers](), [ai-models](), [deployment](), [engineering](), [hardware](), [physics]() ### AI overview Mistral introduces physics AI as a new class of frontier AI models for engineering. The article explains how these models can accelerate physical-system analysis, expand design-space exploration, preserve physics insight during product operation, and support AI-native industrial engineering across manufacturing, aviation, energy, and hardware development. ### Source excerpt A new class of AI models that predict the behavior of physical systems, powering the engineers and hardware products of tomorrow. ## Physics AI research that's shaping the industry. DevFeed: [Physics AI research that's shaping the industry.]() Original publisher: [Read original article]() Published: 2026-05-27T12:00:05Z Content type: news Language: en Sources: [Mistral AI Blog]() Topics: [AI Research](), [Physics-guided deep learning](), [Artificial Intelligence](), [data](), [datasets](), [AI Foundation Models](), [GPU](), [Machine learning](), [Transformer]() Tags: [ai](), [ai-research](), [data](), [datasets](), [deep-learning](), [design](), [energy](), [engineering](), [gpu](), [industry](), [machine-learning](), [models](), [performance](), [physics](), [real-time](), [research]() ### AI overview Mistral describes its acquisition of Emmi AI and its focus on Physics AI for industrial engineering. The article surveys published work on CFD, neural surrogates, foundation models, datasets, plasma turbulence, and real-time industrial simulation across aerospace, automotive, semiconductors, and energy. ### Source excerpt Published breakthroughs pushing the state of the art. ## Emmi joins Mistral to accelerate the AI-native industry DevFeed: [Emmi joins Mistral to accelerate the AI-native industry]() Original publisher: [Read original article]() Published: 2026-05-23T12:00:26Z Content type: news Language: en Sources: [Mistral AI Blog]() Topics: [Artificial Intelligence](), [Simulation and Design](), [AI Development]() Tags: [acquisition](), [ai](), [automotive](), [design](), [engineering](), [industry](), [manufacturing](), [physics](), [platform](), [real-time](), [science]() ### AI overview Mistral AI announces a definitive agreement to acquire Emmi AI, aiming to strengthen its industrial AI offering through Physics AI, engineering models, real-time simulations, digital twins, and AI agents that use existing engineering tools. The deal is intended to support engineering, manufacturing, and high-stakes industrial sectors. ### Source excerpt The most powerful AI platform for enterprises. Customize, fine-tune, and deploy AI assistants, autonomous agents, and multimodal AI with open models. ## GPT-5.2 derives a new result in theoretical physics DevFeed: [GPT-5.2 derives a new result in theoretical physics]() Original publisher: [Read original article]() Published: 2026-02-13T11:00:00Z Content type: news Language: en Sources: [OpenAI News]() Topics: [AI Chat]() Tags: [gpt](), [openai](), [particle-physics](), [physics](), [publication](), [quantum-mechanics](), [research]() ### AI overview GPT-5.2 proposed a formula for a gluon scattering amplitude in a specific half-collinear momentum regime. The article says the result was later proved and verified in a new preprint. ### Source excerpt A new preprint shows GPT-5.2 proposing a new formula for a gluon amplitude, later formally proved and verified by OpenAI and academic collaborators. ## Gemini 3 Deep Think: Advancing science, research and engineering DevFeed: [Gemini 3 Deep Think: Advancing science, research and engineering]() Original publisher: [Read original article]() Author: The Deep Think team Published: 2026-02-12T16:15:09Z Content type: article Language: en Sources: [Google DeepMind News]() Topics: [Google AI](), [Artificial Intelligence](), [Benchmark](), [API](), [Programming](), [Math and Logic]() Tags: [ai](), [api](), [benchmark](), [benchmarks](), [chemistry](), [gemini](), [google](), [math](), [none](), [physics](), [programming](), [reasoning]() ### AI overview Google DeepMind announces an upgraded Gemini 3 Deep Think reasoning mode for challenging science, research, and engineering problems. The article describes its availability in the Gemini app and Gemini API, along with reported results across mathematical, programming, chemistry, physics, and other academic benchmarks. ### Source excerpt Our most specialized reasoning mode is now updated to solve modern science, research and engineering challenges. ## Accelerating Mathematical and Scientific Discovery with Gemini Deep Think DevFeed: [Accelerating Mathematical and Scientific Discovery with Gemini Deep Think]() Original publisher: [Read original article]() Author: Thang Luong; Vahab Mirrokni Published: 2026-02-09T16:12:06Z Content type: article Language: en Sources: [Google DeepMind News]() Topics: [Artificial Intelligence](), [Machine Learning & Artificial Intelligence](), [Combinatorial optimization](), [Algorithms](), [Mathematics](), [Optimization](), [Computer science](), [Training AI Models](), [Streams](), [data](), [Network]() Tags: [ai](), [algorithms](), [combinatorial-optimization](), [data](), [gemini](), [optimization](), [physics](), [research](), [science](), [streams](), [training]() ### AI overview An advanced version of Gemini Deep Think helped researchers resolve long-standing problems across algorithms, machine learning optimization, combinatorial optimization, economics, and physics. The article highlights new counterexamples, mathematical explanations for AI training techniques, an extension of an auction theorem to real-valued bids, and a closed-form solution for integrals involving cosmic-string singularities. ### Source excerpt Research papers point to the growing impact of Deep Think across fields ## Dynamic surface codes open new avenues for quantum error correction DevFeed: [Dynamic surface codes open new avenues for quantum error correction]() Original publisher: [Read original article]() Published: 2026-01-13T17:32:00Z Content type: article Language: en Sources: [The latest research from Google]() Topics: [Google](), [Hardware]() Tags: [google](), [hardware](), [physics](), [quantum](), [research]() ### AI overview Google Research presents an experimental demonstration of dynamic surface codes for quantum error correction. By alternating circuit constructions, the approach offers greater flexibility in gate choice and connectivity while helping address correlated errors, hardware constraints, leakage, and qubit dropouts. ### Source excerpt Quantum ## NVIDIA Cosmos Reason 2 Brings Advanced Reasoning To Physical AI DevFeed: [NVIDIA Cosmos Reason 2 Brings Advanced Reasoning To Physical AI]() Original publisher: [Read original article]() Author: Tsung-Yi Lin; Debraj Sinha Published: 2026-01-05T22:56:51Z Content type: release Language: en Sources: [Hugging Face - Blog]() Topics: [Cosmos](), [Physical AI](), [vlm](), [Artificial Intelligence](), [Video Analytics](), [long-context](), [Deployment](), [data](), [datasets](), [Cloud]() Tags: [agents](), [ai](), [ai-agents](), [analytics](), [cloud](), [cosmos](), [data](), [datasets](), [deployment](), [language-models](), [long-context](), [models](), [nvidia](), [ocr](), [open](), [performance](), [physical-ai](), [physics](), [reasoning](), [vision]() ### AI overview NVIDIA released Cosmos Reason 2, an open reasoning vision-language model for physical AI. The model is designed to help robots and AI agents understand, plan, and act in the physical world, with improved spatio-temporal reasoning, visual perception, OCR, long-context input, and deployment from edge to cloud. ### Source excerpt NVIDIA today released Cosmos Reason 2, the latest advancement in open, reasoning vision language models for physical AI. Cosmos Reason 2 surpasses its previous version in accuracy and tops the Physical AI Bench and Physical Reasoning leaderboards as the #1 open model for visual understanding. Since their introduction, vision-language models have rapidly improved at tasks like object and pattern recognition in images. [Next page]()