AI Briefing#robotics
The latest AI and developer news about #robotics, with the original source and a short summary.
The latest AI and developer news about #robotics, with the original source and a short summary.
OpenAI dissolved its independent robotics team to rely on general-purpose models like GPT-6 Astra for robot control, citing a 74.8% success rate in simulation benchmarks.
The 27B-parameter model is available under Apache 2.0 and outperforms FLUX 3 Action and Cosmos3 Nano on the RoboLab-120 leaderboard.
Four startups use NVIDIA hardware to automate sorting of plastics, recyclables, and textiles, diverting thousands of tons from landfills.
NVIDIA VFlash reduces latency by up to 5.1x for robot and autonomous driving trajectory prediction.
NVIDIA releases GR00T 2, a foundation world-action model that predicts future video and robot actions, enabling zero-shot transfer from human demos and real-time control of Unitree G1 humanoids via the GR00T 2 Turbo variant.
Ai2 (AllenAI) has released ToyBlocks, an experimentation environment capable of analyzing the causes of robotic failures and generating data.
The self-supervised encoder achieves 49.3% tactile-to-video recall at 4K hours, significantly outperforming reconstruction-based baselines.
The open frontier world action model, trained on 100,000 hours of human data, enables skill transfer across robot bodies and ranks among the top three on major robotics benchmarks.
The second-generation accelerator supports 20B to 100B parameter models at 10-30 tokens per second, with projected 6.7x performance gains via Samsung LPDDR5x-PIM.
Korean industry leaders are applying NVIDIA accelerated computing and AI to semiconductor manufacturing, level 4 autonomous driving, and robotic factory automation.
The company plans to use robotic labs to generate data for AI models, creating a feedback loop that accelerates molecular synthesis and testing for pharma clients.
The benchmark separates skill learning from familiar setup success, revealing that current systems fail when block positions or stack heights change.
NVIDIA has released four interactive world simulation models based on Cosmos 3 under the OpenMDW-1.1 license.
The open-source framework pairs automatic data generation with benchmarks showing VLA models generalize poorly while frontier agents solve 30/30 stacking tasks.
The framework uses parallel visual primitive drafting to reduce 3D displacement error by up to 61% and accelerate driving inference by up to 49x compared to autoregressive decoding.
The 16B parameter checkpoints support multi-sensor driving generation and bimanual robot action prediction under the OpenMDW1.1 license.
The open-source uni-robot-agent harness enables Qwen3.8-Flash-Next to reach the performance level of some closed-model robotic agents on real-world manipulation tasks without model weight changes.
The 4GB model targets voice assistants and robotics on RTX, Jetson, and DGX Spark hardware.
Runway announced Praxis-1, a robot policy model pretrained on web videos, and previewed an open-weight release within months.
The model surpasses leading open-weight models on MindCube spatial reasoning and outperforms other world models on FlowBench motion prediction.
Execution speed increased 4.4x and cost decreased 6.2x between July and October 2026, with top models achieving 45-48% task scores.
The model achieves 91.3% accuracy in pinpointing critical video events and supports multi-robot collaboration across various embodiments.
The method, presented at ICML 2026, generates observation and action models via LLMs to reconfigure the agent-environment interface, outperforming human-designed baselines in locomotion and manipulation tasks.
The expanded platform includes five cloud-native data blueprints, support for fine-tuning open models like GR00T and QWEN, and Cosmos 3 Nano variants that enable up to 2x faster inference.