NVIDIA at CVPR 2026

June 3–7 | Denver, Colorado

Advancing the frontiers of computer vision, generative AI, robotics, and physical intelligence.

59

Accepted Papers

39

Workshops

12

Research Areas

42

Open Code Releases


On-Site Events

Connect with NVIDIA researchers and engineers in Denver.

NVIDIA Researcher Celebration

Jun 3  ·  6–9 PM  ·  Invite Only

Join NVIDIA researchers for an evening of conversation, demos, and community. Open to conference attendees with RSVP.

NVIDIA Booth—Live Demos

Jun 3–7  ·  Colorado Conv. Ctr.

Experience hands-on demos of our latest research in physical AI, generative vision, autonomous driving, and robotics.

Launching Legends

Jun 6  ·  12–2 PM  ·  Le Meridien Denver

An exclusive evening for AI entrepreneurs and startup founders. Pitch, connect, and celebrate the next wave of AI-powered companies.

Alpamayo Summit

Jun 3  · 4- 8 PM ·   Le Meridien Denver

A curated gathering of research leaders and industry pioneers exploring the future of AI, computing, and scientific discovery.


Research Papers and Workshops

Browse 59 accepted papers, invited talks, tutorials, and workshops by research area or type.


Code, Models, Demos

Start building with open research from CVPR 2026—download code, pretrained weights, and datasets.

Physical AI

NitroGen: Open Foundation Model for Generalist Gaming Agents

Open foundation model for generalist gaming agents with behavior cloning. Includes model weights, training code, and evaluation benchmarks.

Open source

Foundation

Behavior cloning

Physical AI

SAGE: Scalable Agentic 3D Scene Generation

Scalable pipeline for generating diverse 3D scenes for embodied AI training. Code, pretrained models, and generated scene datasets available.

3d generation

Embodied AI

Synthetic data

Robotics

Fast-FoundationStereo: Real-Time Zero-Shot Stereo

Zero-shot stereo depth estimation that runs in real time. Pretrained checkpoints and inference code for robotics and AV applications.

Stereo

Real-time

Robotics

Generative AI

DuetGen: Interleaved Multimodal Generation

Unified model for interleaved text, image, and video generation. Weights, training recipe, and demo notebooks released.

Multimodal

Image gen

Video gen


Careers

Build What’s Next

Join the team behind 50 papers at CVPR. We’re hiring researchers and engineers across AI, robotics, graphics, and more.


Explore NVIDIA Research

Go beyond the conference—discover the labs, tools, and open research powering these breakthroughs.

Research Labs

Explore cutting-edge work from NVIDIA’s research labs, spanning AI, graphics, robotics, and autonomous systems. Access code, publications, and interactive demos.

Publications

Browse the full catalog of peer-reviewed NVIDIA papers across all conferences, journals, and preprints—searchable by topic, year, and research area.

AI Playground

Try interactive demos of our latest research in generative AI, vision, speech, and more—no setup required. See the science in action before diving into the code.

Sign up to receive the latest news from NVIDIA.