Repo of VisionWeave: Weaving Elastic Visual Representations as a Native Capability of MLLMs

Code to pretrain, fine-tune, and evaluate DreamZero and run sim & real-world evals

Official code of Motus: A Unified Latent Action World Model

Official Project Page for Recurrent Looped Transformer (RLT) (https://arxiv.org/abs/2610.07591)

An all-in-one VLA engineering platform for embodied AI — from data to real-robot deployment.

Official implementation of AURORA-LM: Autoencoding Unified Representation for Continuous-Latent Diffusion Language Modeling.

Self-Supervised Latent Motion Priors for Physical Realism in Video Generation

Official implementation of "What Makes World Action Models Generalize? An Empirical Study of Test-Time Future Modeling"

A Fully Continuous Multimodal Flow Framework For Flexible Multimodal Sequence Modeling.

InternW0-Δ : A World Action Model Bridging Predictive Dynamics and Actions with 20K+ Hours of Open Data

Executable, measurable, and reproducible AI4AI toward recursive self-improvement. Home of OpenMLE and Frontis-MA1.

Cuti Harness: An open-source, plugin-based video world harness

GPT-6 Astra for embodied AI and robotics.

GPT-Policy: In-Context Robot Learning with VLM Agents

This is the official repository for paper "Adaptive Vision-Language Grasping via Composable Foundation Priors and Generalizable Grasp Synthesis".

RPent: Agentic Infrastructure for the Physical World

MiniCPM5: SOTA on-device LLMs, small yet powerful.

Pretraining and inference code for a large-scale depth-recurrent language model

Official repository for "OpenWAM: An Open, Modular Exploration Towards Systematic World–Action Model Pretraining".

[CoRL2026] MotionVLA: Injecting Geometric Motion into Vision-Language-Action Model

S-Space: Exploring Spatial Workspace in Multimodal Models

[ICLR 2026] Official implementation for What matters for Representation Alignment: Global Information or Spatial Structure?

[NeurIPS 2026] Beyond Data Scaling: Representation-Centric Continued Pre-training for Vision-Language-Action Models

Open data and scalable training for long-horizon video world models.

Open-source framework for the research and development of foundation models.

一个简单易上手的 Obsidian 研究卡片管理器，支持 Markdown 卡片、关系画布、箭头标 签与颜色、框选多选和状态看板。