MBZUAI’s IFM releases open 29B PAN 2 world model
Apache 2.0 weights and code ship; IFM reports 34% full-plan completion, ahead of MiniMax-H3 on multi-step tasks.
1
Stories per day · 30 days
24 RSS
Relatedembodied-ai25vision-language-action24vla21world-models21simulation20reinforcement-learning19libero15manipulation13benchmark12nvidia12physical-ai12vlm9dexterous-manipulation7sim-to-real7world-model7benchmarks6imitation-learning6vision-language6
398 stories, newest first · one card per story
Stories tagged #robotics
NVIDIA profiles startups using vision AI to sort mixed waste
Four Inception companies run NVIDIA edge GPUs to identify plastics, packaging and textiles on fast conveyor belts.
1
Amazon launches Ring indoor drone and wheeled business patrol robot
Both robots map interiors and use cloud processing; Ring cites a scheme it says later deletes encryption keys.
NVIDIA introduces GR00T 2 robot world action model at CoRL
NVIDIA says GR00T 2 hit 71% average success across 50 tasks with one demonstration and 95% with five.
One VLA policy places 9th in the 2026 BEHAVIOR Challenge
A single π0.5-based checkpoint scores 0.392 Q on 100 tasks; GPT-6 Astra averages 0.639 Q on 14 routed tasks.
DeepMind unveils Gemini Robotics 2.5 and ASIMOV-jailbreak benchmark
Apollo 2 robots kitted parts in Mercedes factories at 96% success; the open jailbreak set has 1,500 prompts.
Latent Barrier Steering adds hierarchical safety to generative planners
On navigation and robot end-effector tasks, LBS kept safety while improving goal reaching versus path-level filters.
ExRL adapts long-chunk robot policies by learning execution length
A lightweight off-policy RL adapter predicts how long to run each action chunk, cutting the adaptation space versus optimizing the full chunk.
DexForesight lifts dexterous policies with implicit spatial foresight
On 11 DexJoCo tasks success rises from 49.4% to 60.5%; three real-robot tasks reach 58.3% versus 50.0%.
Memory VLAs treat pauses as progress via temporal shortcuts
Brief pauses drop counting success from 47 to 16 of 50 and from 44 to 10; event-indexed memory recovers half or more without retraining.
NVIDIA Announces Investments in Korean AI Startups
NVIDIA will invest in Korean startups Point2, Motif Technologies, RLWRLD and Upstage following the MSIT-NVIDIA forum.
NVIDIA advances Korea physical AI ecosystem
NVIDIA helps Korean manufacturers apply AI and accelerated computing to semiconductors and robotics.
SWAM studies how action learning affects language-video alignment
Authors report 88.10% easy and 61.56% hard RoboTwin success, and 40.00% versus 23.33% on three real tasks.
RoboScript learns manipulation tasks in context from one demonstration
Authors report 83.4% mean success on all 50 RoboTwin 2.0 tasks without gradient updates, over 20 points above the strongest trained policy.
Embodied First Exam tests coding agents on 1,156 robot tasks
GPT-6 Astra in Codex, strongest of eight agents, succeeds on 34% of 456 core tasks.
ArmedAgent finds safety risks in coding agents that control robot arms
The benchmark covers 36 tasks in five scenarios, including indirect prompt injection that can move from digital resources into physical actions.
Physicl raises $10M to simulate realistic worlds for robots
Physicl repurposes its 3D rendering tech to generate high-fidelity simulations for training physical AI models.
A2-MEAS embodied framework reports BEHAVIOR and real-robot scores
Authors report 0.6411 Q-score on 100 public tasks and 57.92% success on 12 real-robot tasks versus 26.67% for pi0.5.
Agility goes public via SPAC, raises $465 million
Agility plans to demonstrate Digit 5 at CES 2027 and has contracts from 30 customers.
1
Dynamic watermarking detects replay attacks in vision controllers
RL policy balances watermark energy against control distortion on UR5 robot.
SLA Model Scales Spatial Understanding via Continual Pre-training
SLA optimizes action prediction and spatial alignment from frozen 3D perception for robotic manipulation.
VFlash Enables Fast Visual Planning via Flash Primitive Decoding
VFlash predicts visual primitives jointly for 25x faster generation in robot trajectory tasks.
NEO-VLA: Native Vision-Language-Action Framework Outperforms Encoders
Native vision-language-action models achieve stronger robotic control by evolving visual representations jointly with language.
PyRUA-DataFarm automates robot data collection
PyRUA-DataFarm cuts robot data costs to $21 per task using an LLM agent in simulation.