Sunday Robotics released the foundation model ACT-1, claiming it was trained without any robot manipulation data; instead, human motions were captured with a Skill Capture Glove and converted into robot data via Skill Transform with a 90% success rate.
Physical Intelligence has released a new model, π0.6, which can learn from its own experience through reinforcement learning (RL). Compared with a baseline model not trained with RL, π0.6 can more than double throughput and can perform real-world tasks, including making espresso drinks, folding various garments, and assembling cardboard boxes.
Why it matters: The original text provides a comparison showing more than double the throughput of a baseline model not trained with reinforcement learning, as well as the scope of three types of real-world tasks.
Physical Intelligence released the VLA model π*0.6. After training with its Recap (RL with Experience & Corrections via Advantage-conditioned Policies) method, throughput on some of the hardest tasks more than doubled, and the failure rate can be reduced to half of the original or lower.
Why it matters: The original article explains how compounding errors arise in VLA under imitation learning and presents a training approach that uses expert corrections and autonomous experience for reinforcement learning.
A team led by Nitin J. Sanket, a professor at Worcester Polytechnic Institute (WPI), has developed palm-sized flying robots that mimic bats' use of ultrasound, aimed at search-and-rescue missions. AI software filters noise from the ultrasound signals, letting the robots detect obstacles within a 2-meter radius. The team chose low-power ultrasonic sensors commonly used in automatic faucets, and to suppress interference from propeller noise, designed 3D-printed structures placed in front of the robot to modulate sound waves, similar to the function of a bat's nose, ears, and mouth tissues. The robot currently works; the next step is to increase its flight speed.
Dexterity proposes "Transactable World Models," which implement spatial understanding as a shared operator subject to physical constraints; multiple asynchronous skill agents interact with it transactionally, providing uncertainty bounds, rollback, interpretability, and real-time consistency.
We're excited to share our latest progress on DYNA-1 pretraining! 🤖 The foundation model can now perform a variety of dexterous tasks (folding clothes, parcel sorting, ...) without any post-training, even in environments it has never seen. This powerful foundation also supports extremely efficient fine-tuning: with only about 1 hour of data, it can reach about 100% success on challenging new tasks! 🤯 Check out two tasks it has mastered: stacking cups and cutting celery, performed repeatedly with zero failures. 👇
Holiday Robotics has unveiled Friday, an industrial humanoid robot whose hands use a 500-gram, 20-DOF hydraulic ultralight design, with about 80% of their surface covered by self-made tactile sensors that can sense a force of 0.01N (about 1g), as well as direction and torque.
At CeMAT ASIA 2025, Sainade made its debut showcasing the upgraded linkage of iLoabot-M and iLoabot-X, with a live on-site demonstration of the full process of "loading and unloading—depalletizing and palletizing—transfer," in which the two machines worked in coordination to complete an unmanned closed-loop operation.
🎃 Halloween is coming. Our hardworking team is lining up for candy, of course, served by Dynasaur! The DYNA VLA model now has robust agentic reasoning capabilities and can serve any combination and quantity of candy! Given that the number of possible combinations is exponential, pure imitation learning will not work. No video editing. Continuous live footage, as always 🤖 DYNA wishes everyone a happy Halloween! 🍬
Diligent Robotics announced the launch of Moxi 2.0, which the company says delivers 10 times the compute of Moxi 1.0 and is built on the NVIDIA Thor platform. The platform was trained on three years of real-world hospital deployment data and more than 1.25 million deliveries; Moxi currently delivers medications and lab samples for nurses and pharmacy staff at more than 25 hospitals across the U.S.
Running the GR00t N1.5 model to pour matcha. Getting ready for the @seeedstudio hackathon @NVIDIARobotics @VectorWang2
Xinyichang has officially released the first-generation HOSON-Robot embodied intelligent humanoid robot and will proceed with subsequent small-batch mass production. The robot features the self-developed RMC100 "cerebellum" motion controller, supporting synchronized drive of 64 joint motors with synchronization error below 100ns and a control frequency of 4kHz; its model adopts a hybrid intelligence architecture that integrates a VLA end-to-end model with a fast-slow dual-system hierarchical model, and its self-developed dexterous hand and lightweight joint modules have completed preliminary adaptation for industrial scenarios.
Tashi Intelligence has released World In Your Hands (WIYH), described as the world's first large-scale real-world embodied VLTA (vision-language-tactile-action) multimodal dataset, and plans to share it with the industry in December 2025.
Tianlian Robotics has published a video imitation learning training pipeline: extract reference motions with GVHMR/TRAM (or use motion capture datasets such as AMASS), then after human-to-robot joint mapping and training a motion orientation model, use PPO reinforcement learning in simulation with curriculum learning to progress from static poses to long-sequence complex motions, then optimize through domain randomization before deploying to the real robot, with fine-tuning via residual reinforcement learning or adaptive gains. The company says it has achieved key breakthroughs in dynamic force-control precision and stability, and that motions can also be reproduced in real robotic arm operation.
Sainade showcased its full iLoabot lineup of three robots, iLoabot-M, iLoabot-P, and iLoabot-X, at the 2025 China International Industry Fair (CIIF), with live demonstrations of truck loading and unloading and slope climbing and obstacle crossing. iLoabot-X is its first heavy-duty dual-arm humanoid robot, built on Sainade's self-developed VLA, the logistics industry's first vertical large model, with an initial positioning time of 5 minutes and full-machine production in 30 minutes; the company says the iLoabot series has already been deployed at scale by leading enterprises in tobacco, beverages, and express delivery.
KEENON Robotics has officially released KOM 2.0 (KEENON Operator Model 2.0), which it calls the world's first VLA (vision-language-action) model for the service industry. The model uses a dual-system architecture with the K-Mind thinking system and the K-Act action system, is trained on hundreds of millions of real service interaction data points, and supports millisecond-level real-time decision-making.
The paper "Generative Visual Foresight Meets Task-Agnostic Pose Estimation in Robotic Table-top Manipulation," a collaboration between SUSTech CLEAR Lab, LimX Dynamics, and the University of Hong Kong, has been accepted to CoRL 2025.
Why it matters: This research presents a concrete technical approach and experimental figures for learning manipulation from human video, helping readers understand the feasibility of training robot manipulation with video data.
Figure announced two advances in Project Go-Big: a partnership with Brookfield to build a large-scale humanoid pretraining dataset, with Brookfield holding over 100,000 homes, 500 million square feet of commercial office space, and 160 million square feet of logistics space; and Helix, after training on 100% first-person human video, responding to natural-language commands such as "go to the fridge" without any robot demonstrations, navigating cluttered real home environments.
Why it matters: The original gives the training method by which Helix learned navigation using only human video and the scale details of the Brookfield data partnership, offering a view of actual progress on the humanoid pretraining data approach.
Figure AI announced that its VLA (vision-language-action) model Helix learned to load a dishwasher using only new data, without changing the algorithm or adding special engineering. Helix can separate stacked plates and load them in order, pick up a glass with one hand, reorient it, and carefully place it with the other hand; it can also handle cluttered starting states and recover from grasp failures or collisions. The same model previously completed towel-folding and package-sorting tasks, and the company says this reflects the direction of a single general system learning new capabilities incrementally with data.
LimX Dynamics has introduced LimX DreamActor, an embodied intelligence training paradigm that for the first time deeply integrates Real2Sim2Real, imitation learning, and real-robot reinforcement learning into a full pipeline of "real-world environment data collection—simulation data pretraining—real-robot data post-training." The system can complete multi-view data collection using only a consumer-grade camera; through 3D reconstruction and physical parameter alignment it builds a highly realistic simulation scene, then uses a small amount of real-robot reinforcement learning to complete cross-domain transfer, aiming to improve sample efficiency, convergence speed, and deployment stability.
The 2025 World Robot Conference concluded on August 12. Estun Cuzhuo exhibited the CODROID02 humanoid robot and the Co-Palletizer 20 lifting palletizing system. CODROID02 stands 170 cm tall, weighs 70 kg, has a single-arm payload of 5 kg, and has 31 degrees of freedom across its body (excluding dexterous hands). Its joint modules are fully self-developed, and it uses a fast-slow system architecture combining a vision-language large model with a flow matching strategy. On site, it demonstrated part grasping and barcode scanning.
Figure AI's VLA model Helix fully autonomously folds towels, which the company says is the first time a multi-fingered humanoid robot has completed this task using an end-to-end neural network. With the architecture and training hyperparameters unchanged and only a new dataset added, it can pick items from a mixed towel pile, adjust its folding strategy based on the initial shape, correct multiple grasping errors, and untangle twisted towels using fine motions such as tracing edges with its thumb and pinching corners.
Why it matters: The same model completes the towel-folding task with no architecture changes, relying only on added data, demonstrating a generalization path from logistics to deformable object manipulation.
Sainade's heavy-duty dual-arm autonomous working robot iLoabot-X made its debut at the World Robot Conference (WRC), demonstrating on-site dual-arm collaborative handling of a cumulative 40 kg of cargo to complete transfer, as well as autonomous recognition of cage cart positions and path planning to push them. Its autonomous working capability comes from Sainade's self-developed multimodal VLA vertical large model, the first in the logistics industry, which parses operational environment details through multimodal perception. The same series, iLoabot-M, has already been running stably in the tobacco, alcoholic beverage, and express logistics sectors, with the company claiming a 300% efficiency improvement over traditional solutions.
SAIC Beijing and LimX Dynamics signed a strategic cooperation memorandum in Beijing, announcing the establishment of a joint embodied intelligence laboratory to conduct embodied robot algorithm R&D, system development, and toolchain deployment around automotive scenarios. LimX Dynamics will provide underlying technologies including robot bodies, whole-body motion control algorithms, VLA model frameworks, data engines, and toolchain platforms. The two parties will carry out humanoid robot data collection, training, and evaluation in automotive manufacturing scenarios, and gradually form a complete R&D process from data collection and model training to system testing.
On June 11, Estun Cobot officially launched its second-generation humanoid robot, Codroid 02, at the 2025 Intelligent Robot Development Conference. It stands 170 cm tall, weighs 70 kg, has a single-arm payload of 5 kg, and has 31 degrees of freedom across its body (excluding dexterous hands), with all core components independently developed in-house.
Figure AI released progress on Helix's capabilities after three months of deployment in a logistics environment: per-item processing time fell from about 5.0 seconds to 4.05 seconds, barcode scanning success rate rose from about 70% to about 95%, and the system gained the ability to handle flexible packages such as poly bags and flat envelopes.
Why it matters: The original article provides the data scale and ablation experiment details for each architecture module, allowing readers to understand the specific sources of the logistics performance improvement.
This article explains NVIDIA's open-source humanoid robot foundation model Isaac GR00T N1, unveiled at the March 2025 GTC conference, along with the accompanying GR00T Blueprint synthetic data toolchain and Newton physics engine.
Figure AI released an end-to-end reinforcement learning walking controller that lets the Figure 02 humanoid robot learn a human-like gait in simulation by imitating human reference trajectories through rewards, including heel strike, toe push-off, and arm swing. Training runs in parallel across thousands of robots in GPU-accelerated physics simulation, collecting years of simulated data in hours, and combines domain randomization with kHz-level torque feedback control to achieve zero-shot transfer to the real robot; 10 Figure 02 units run the same neural network without individual tuning.
Why it matters: The original text provides details on the training method and sim-to-real transfer for Figure 02's reinforcement-learning-based human-like walking, offering insight into its technical approach.
FieldAI claims to have developed the first risk-bounded Field Foundation Models (FFM) robot foundation model, which enables multi-form robots to operate autonomously in hazardous, off-road, and other unstructured industrial environments without prior maps, GPS, or cloud communication, and has already been deployed at hundreds of sites.
USC and NVIDIA jointly introduced the ProgPrompt model, which uses a language model to plan robot tasks in the form of Python-style code, making the generated plans reusable across environments, robot capabilities, and tasks.