Cosmos3

In the model registry

Timeline

1. Oct 20, 2026 · Model release · 2 sources
   NVIDIA Cosmos3-H-Dreams: Bringing Clinical Surgery to Life Through Generative Simulation
   Our first Cosmos-H-Dreams release, built on Cosmos-Predict2.5, brought interactive generative simulation to dVRK tabletop suturing, a foundational proving ground for surgical robotics.
2. Oct 20, 2026 · Research paper · 1 source
   Industrial ReMEmbR: Scaling Spatiotemporal Memory to Robot Fleets
   We present IndustrialRemembr, a shared spatiotemporal memory system that extends NVIDIA ReMEmbR to robot fleets and reduces server-side computation by generating scene descriptions only on demand.
3. Oct 19, 2026 · Product / feature launch · 1 source
   Amazon SageMaker JumpStart: 2026 year-to-date in review
   Amazon SageMaker JumpStart is a machine learning (ML) hub within Amazon SageMaker AI that provides access to a curated catalog of AI models, solution templates, and example notebooks.
4. Oct 14, 2026 · Open-source release · 1 source
   NVIDIA/TensorRT-LLM v1.4.0rc0
   DeepSeek MLA models on SM120/SM121 GPUs (for example RTX PRO 6000 Blackwell) may crash with an illegal memory access when CUDA graphs are captured with BF16 BMM.
5. Oct 13, 2026 · Research paper · 1 source
   AffectWorld: Benchmarking Emotion-Driven Future World Prediction
   However, existing world model benchmarks primarily assess physical plausibility and perceptual quality, leaving emotion-driven world evolution largely untested.We introduce AffectWorld, a benchmark that evaluates the chain Emotion-Action-World State through three tasks: World-State Consistency, Emotion-Action Consistency, and Causal Consistency.
6. Oct 13, 2026 · Research paper · 1 source
   The Human--VLA Communication Gap: What Humans Leave Implicit and What Policies Cannot Recover
   We study this human--VLA communication gap by separating an information gap, where task semantics, grounding, or event structure remain unresolved, from a residual execution gap, where even a fully specified task depends on how it is organized and invoked.
7. Oct 9, 2026 · Research paper · 2 sources
   Learning Transferable World-Action Models from Task-Paired Human Videos
   We introduce WATCH, a cross-video pre-training that learns from task-paired human demonstrations collected across diverse environments.
8. Sep 29, 2026 · Open-source release · 1 source
   NVIDIA/TensorRT-LLM v1.3.0rc29
   Expose Nemotron-H vision-language LoRA configuration for supported inference paths #19151

Often appears with