Skip to content
Only Models#Grasping and manipulation
  1. 量子位88

    Ant Group's Lingbo open-sources LingBot-Touch, a tactile pretraining foundation model

    Ant Group's Lingbo has open-sourced LingBot-Touch, a tactile pretraining foundation model that uniformly encodes magnetic, visuotactile, and piezoresistive tactile signals into tactile tokens, which are then passed to the action model. Team experiments showed that when π0.5 was directly fed raw tactile readings on a zipper-pulling task, its success rate dropped from 39% to 15%; after connecting LingBot-Touch, it rose to 59%.

    Why it matters: The original provides a controlled experiment showing that feeding raw tactile signals into the action model instead lowers the success rate, along with the improvement after unifying tactile tokens, which can serve as a reference for tactile representation design.

  1. Beijing Humanoid Robot Innovation Center70

    Beijing Humanoid releases Hi-WBC general whole-body control model on "Huisikaiwu" platform

    Beijing Humanoid released the Hi-WBC general whole-body control model on the "Huisikaiwu" platform on November 2. The model combines human motion references with real physical constraints, enabling robots to adjust how they execute tasks as payload, external forces, and terrain change. It uses a continual skill expansion framework based on MoE and policy distillation, Contact-aware Learning that incorporates contact force prediction, and Perceptive Control that introduces terrain awareness.

    Why it matters: The original text discloses Hi-WBC's technical approach to handling payload, external force, and terrain constraints beyond motion tracking, offering insight into how whole-body control is moving from demonstrations toward real work.

  2. 机器之心 · 微信公众号80

    LiberAI releases world action model Liber-0, publishes technical report and online simulator

    LiberAI, founded by Liu Songming, has released Liber-0, a new-generation world action model that uses three expert streams for language, vision, and action along with Multimodal Attention to jointly predict future frames and actions.

    Why it matters: The technical report uses controlled experiments to break down the contributions of human data, action supervision, and embodiment diversity to dexterous manipulation, which can inform judgments about data-mix strategies.

  1. 未来不远 · 微信公众号76

    Future Not Far releases Self-Evolving WAM self-evolving world-action model

    Future Not Far released the Self-Evolving WAM self-evolving world-action model, integrating joint world-action modeling, interactive experience learning, and training-effect feedback into a single system. In update attempts from round 0 to round 30, all four groups of household tasks showed gains in autonomous full-completion success rates: folding clothes rose from 54.6% to 96.8%, hanging clothes from 46.2% to 91.2%, making pasta from 42.7% to 89.5%, and stir-fried pork with chili peppers from 38.2% to 68.9%.

    Why it matters: The original text gives the success-rate changes for four groups of household tasks from round 0 to round 30 updates, showing the practical benefits of the self-evolving training loop.

  1. 机器之心 · 微信公众号82

    X-Era AI opensources 4D world action model E0.5, real-robot success rate surpasses π0.5

    X-Era AI has open-sourced its native 4D world action model E0.5, achieving overall average success rates of 85.0% and 88.3% on the AgileX Piper dual-arm robot and the Zerith H1 humanoid robot, respectively, 13.3 and 26.7 percentage points higher than π0.5.

    Why it matters: The original text provides real-robot success rate comparisons between E0.5 and π0.5 on dual-arm and humanoid platforms, which can be used to assess the manipulation generalization performance of the 3D geometric backbone approach.

  1. Xbotics具身智能实验室 · 微信公众号76

    Shanghai AI Lab evaluates GPT-6-Astra's capability boundaries as a robot policy on EBench

    Shanghai AI Laboratory's embodied evaluation team used EBench to evaluate GPT-6-Astra as a robot policy directly, without task-specific post-training, across 26 tasks and 510 total executions. Its average task success rate of 46.7% ranked third, and its average task score of 0.654 ranked second; OpenWAM-α led with a 55.3% success rate.

    Why it matters: The benchmark quantifies the capability gap between general large models and post-trained policies, helping readers judge which task lengths and manipulation precisions each approach suits.

  1. 至简动力 · 微信公众号72

    Simplexity Robotics releases and open-sources world-action model SimpleWAM, with 6B model achieving on-device inference as low as 42.31ms

    Simplexity Robotics has launched and open-sourced the world-action model SimpleWAM; the 6B-parameter version reduces inference latency on NVIDIA ThorX (T5000) from 181.23ms to 42.31ms, enabling high-frequency closed-loop operation on-device at nearly 24Hz.

    Why it matters: The original article presents the on-device optimization path from 181ms to 42ms and four architectural designs, which can be transferred to understand how world-action models can meet high-frequency closed-loop control requirements.

  1. Figure AI79

    Figure AI releases Helix 2.5, achieving zero-shot full-body autonomy in 30 unfamiliar homes

    Figure AI released Helix 2.5, which completed three long-horizon full-body behaviors—tidying a living room, folding towels, and making a bed—zero-shot in 30 Bay Area homes where no data had ever been collected, without collecting data or fine-tuning in the evaluation environments.

    Why it matters: The original article provides controlled comparison data on Index pretraining for zero-shot generalization and evaluation details, helping readers understand the actual level of humanoid robot pretraining transfer.

  1. Skild AI76

    Skild AI releases robot foundation model S1, achieving in-context learning through video demonstrations

    Skild AI released the robot foundation model S1, which uses video demonstrations as task prompts and can execute manipulation tasks not seen during training without fine-tuning, with long-horizon tasks lasting up to about 10 minutes. At a pretraining data scale of 100,000 hours, the ICL model achieved a 66% success rate on unseen tasks, compared with 9% for the language-prompted VLA; a single demonstration was equivalent to about 380 post-training demonstrations, and the planting task took 11 minutes from recording the demonstration to autonomous execution. S1 was trained on NVIDIA AI infrastructure and has already been put into use with commercial partners.

    Why it matters: The original article provides comparison figures for ICL and VLA at the same data scale and a deployment timeline, allowing readers to assess the feasibility of the in-context learning approach for manipulation tasks.

  1. Sunday Robotics78

    Sunday Robotics previews ACT-2 model with 99.1% zero-shot success rate in clothes folding

    Sunday Robotics previewed its robot model ACT-2, which achieved a 99.1% (±0.3%) zero-shot success rate in clothes folding across diverse unseen home environments with zero per-home adaptation, covering 9 clothing categories across 785 autonomous attempts, with an average folding quality of 4.72/5.

    Why it matters: The official disclosure provides detailed quantitative data on pretraining scale and generalization gaps, allowing readers to understand how a small amount of demonstration data transfers to unseen environments.

  1. LimX Dynamics61

    LimX Dynamics releases humanoid VLA model V³-0 alongside COSA 0.5

    LimX Dynamics released V³-0 alongside COSA 0.5, a humanoid VLA technology stack in its VLA-tools layer that supports mobile whole-body manipulation, running on the 43-DOF LimX humanoid robot Oli.

    Why it matters: The original text explains the architectural composition of V³-0 and its human-in-the-loop real-robot reinforcement learning loop, which lets readers understand where the mobile whole-body manipulation capability comes from.

  1. PsiBot/灵初智能81

    Lingchu Intelligence and PKU-Lingchu Joint Lab Release and Open-Source Dual Dexterous Hand General Manipulation Model and Full-Stack System EgoSteer

    Lingchu Intelligence and the PKU-Lingchu Joint Lab released and open-sourced the dual dexterous hand general manipulation model and full-stack system EgoSteer, comprising the data processing pipeline EgoSmith, the robotics technology stack Robot-Stack, and a world-model-enhanced VLA model.

    Why it matters: The original text presents a complete chain from human-hand data processing and real-robot correction to model architecture and experimental figures, allowing readers to assess its reproducibility value.

  1. X Square Robot/自变量机器人67

    X Square Robot releases cross-modal embodied action tokenizer X-Tokenizer, with 13.5% gain in multimodal alignment and 8.25% gain in long-horizon tasks

    X Square Robot has released X-Tokenizer, a cross-modal embodied action tokenizer built on a semantic residual quantization (SRQ) architecture and pretrained on 17 robotic arm series and 2.4 million trajectories (including 2 billion action frames). Compared with FAST, it improves multimodal alignment by 13.5% and long-horizon task performance by 8.25%, and scores 82.8 on RoboTwin 2.0; in real-robot tests it outperforms comparison methods such as FAST and 4-level RVQ.

    Why it matters: The original text presents X-Tokenizer's method design and comparative experimental data, letting readers understand the action tokenizer's multimodal alignment approach and its reusable effects.

  1. HUMANOID63

    KinetIQ Ascend uses real-world reinforcement learning to boost humanoid robot manipulation throughput and success rate

    HUMANOID has launched KinetIQ Ascend, an end-to-end visual reinforcement learning method that operates under production deployment conditions, extending its KinetIQ framework to let robots improve manipulation skills through trial and error rather than relying solely on human demonstration; the company says this is the first such method publicly demonstrated on production-line humanoid robot manipulation.

    Why it matters: The original text provides specific throughput and success-rate figures for three industrial tasks and notes a pattern of predictable performance improvement with training duration, making it useful for assessing the practical value of real-world reinforcement learning on humanoid robots.

  1. Figure AI62

    Figure AI shows two Helix-02 humanoid robots collaborating to tidy a bedroom

    Figure AI demonstrated two humanoid robots equipped with Helix-02 collaborating to tidy a bedroom within two minutes, completing tasks such as opening a door, hanging clothes, putting away headphones, closing a book, taking out trash, pushing a chair under a desk, and making a bed together. The two robots run the same learned vision-language-action (VLA) policy, with no shared planner, message passing, or central coordinator; each infers the other's intent from its actions via its own cameras. The company says that to its knowledge this is the first demonstration of multi-humanoid collaborative mobile manipulation completed by a single learned neural network directly from pixels to actions.

    Why it matters: The official demonstration shows two humanoid robots driven by the same VLA system collaborating to tidy a bedroom, giving readers insight into an approach to multi-robot collaboration without a shared planner.

  1. PsiBot/灵初智能75

    Lingchu Intelligence releases Psi-R2 and Psi-W0, open-sources 1,000-hour full-modality human hand manipulation dataset

    On April 10, Lingchu Intelligence released Psi-R2, pretrained on human data, and the Action-Conditioned World Model Psi-W0, and open-sourced the first 1,000 hours of a full-modality human hand manipulation dataset, with total data reserves reaching 100,000 hours.

    Why it matters: The original text provides the division of labor between the two new models and the open-source data entry point, so readers can assess the feasibility of using human data for robot training.

  1. Physical Intelligence77

    Physical Intelligence releases steerable general-purpose robot model π0.7, demonstrating compositional generalization

    Physical Intelligence has released the general-purpose robot model π0.7, which it says can perform a variety of dexterous manipulations with speed and robustness comparable to fine-tuned specialized models, and shows preliminary signs of compositional generalization on tasks not seen in its training data.

    Why it matters: The article uses two specific experiments, an air fryer and folding clothes with a UR5e, to show the actual performance of compositional generalization and cross-embodiment transfer, making it easier to judge the gap between general-purpose and specialized models.

  1. PsiBot/灵初智能74

    Lingchu Intelligence Releases Psi-R2 and Psi-W0 Dual Models, Pretrained on 100,000 Hours of Human Data

    Lingchu Intelligence released two models, Psi-R2 and Psi-W0, which it calls the industry's first embodied intelligence models pretrained on human data at the 100,000-hour scale.

    Why it matters: The original article, written by the team itself, breaks down the dual-model architecture and data governance conclusions, offering a reusable path for human data utilization along with experimental evidence.

  1. Physical Intelligence74

    Physical Intelligence proposes a physical intelligence layer and publishes π 0.6 deployment data from Weave and Ultra

    In a blog post, Physical Intelligence argues that robots need a reusable physical intelligence layer, and it published real-world deployment data from partners Weave and Ultra using π 0.6.

    Why it matters: The two deployers provided comparative data on autonomy rates, missed grasps, and throughput, making it easier to judge the practical gains that foundation model iterations bring in real-world settings.

  1. Figure AI82

    Figure AI Releases Helix 02: A Full-Body Autonomous Control Model for Humanoid Robots

    Figure AI has released Helix 02, extending control from the upper body to the entire humanoid robot and enabling continuous, integrated autonomous tasks combining locomotion and manipulation. It adds a new System 0 whole-body controller, trained on over 1,000 hours of human motion data and simulation-based reinforcement learning, which outputs joint commands at 1 kHz via a 10M-parameter network, replacing roughly 109,504 lines of hand-written C++ code; System 1 connects head-mounted vision, palm cameras, fingertip touch sensing, and proprioception, outputting whole-body joint targets at 200 Hz.

    Why it matters: The original text provides details on Helix 02's architectural layers, training data, and demonstration tasks, allowing readers to understand the specific implementation approach for full-body autonomous control.