NVIDIA adds Cosmos tools to train and deploy world models
Original title: NVIDIA Cosmos Opens New Tools to Train, Evaluate and Deploy World Models
New blueprints, Cosmos 3 Nano simulation variants, and NIM microservices claim up to 2x faster inference.
NVIDIA expanded Cosmos with Data Factory blueprints, training and optimization skills, and simulation checkpoints for physical AI and world models. The factory adds five cloud-native blueprints for driving data, image and video augmentation, rare-event video, and defect images, available through GitHub, Azure, Nebius, and CoreWeave. Cosmos Framework support now covers models including GR00T, Alpamayo, Qwen, and Pi0.5, plus Cosmos 3. Customer examples from Porsche, ANYbotics, ADNOC, and others include expected or claimed gains.
Unconfirmed: Customer speed and cost reductions are NVIDIA-reported expectations or claims, not independent measurements.
- Porsche expects up to 90% fewer repeated test drives and 85% shorter cycles.
- MicroAGI distilled a 32-billion-parameter Cosmos 3 reasoner to an 8-billion Nano model.
- Updated Cosmos 3 NIM microservices claim up to 2x faster inference in validated setups.
- Metropolis VSS 3.3 is said to monitor 40% more video streams on the same GPU.
- Vendors
- NVIDIA
From the original article
4 min read at blogs.nvidia.com
To help developers rapidly build world models, which help physical AI systems understand and simulate the real world, NVIDIA has expanded the Cosmos platform with new training tools, simulation checkpoints and Cosmos Data Factory blueprints. Ranking top five on six different open models leaderboards, the NVIDIA Cosmos platform gives world model builders the frontier open […]
By Pranjali Joshi · blogs.nvidia.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.