Code for "FACT: Failure-Aware Causal Training for World‑Action Models"

DeepSeek Harness: Everything is a Plugin.

SCoPE: Sightline-Coordinate Positional Encoding for 3D-Aware Video Generation

LiLa-WAM: Lightweight Latent Reasoning World-Action Model for Robotic Manipulation

Official Code of SWEET: Sparse World Modeling with Image Editing for Embodied Task Execution

Repository associated with paper titled "FOCA: Future-Oriented Conditioning for Data-Efficient Vision-Language-Action Adaptation", presented at ICML 2026.

Training and evaluation code for paper "Learning to Move Before Learning to Do: Task-Agnostic pretraining for VLAs"

[CVPR 2026] Official implementation of "Action-Sketcher: From Reasoning to Action via Visual Sketches for Long-Horizon Robotic Manipulation"

Official Code For VLA-OS.

This repo is the official implementation of "τ0-VLA: a Hierarchical Robot Foundation Model with World-Model-Guided Test-Time Computation".

[CVPR2026] Chain of World: World Model Thinking in Latent Motion

NORA: A Small Open-Sourced Generalist Vision Language Action Model for Embodied Tasks

[Official Repo] JoyAI-Video-Edit: Real-Time Open-Ended Video Editing with Autoregressive Diffusion

Unsupervised Learning of Generalizable Robot Motion from Compact State Representation

ThinkJEPA: Empowering Latent World Models with Large Vision-Language Reasoning Model

Official Implementation of Pi-R2

Towards Versatile Vision-Language-Action Models via Learning Unified Vision-Motion Representations

(ICML2026) Official implementation of VLANeXt.

Hierarchical Sub-Goal Policies for Generalizing Robot Manipulation

INTACT: Isomorphic Intent-to-Action Learning for Search-Free World Models.

Boogu-Image-0.1 is an Apache-2.0 open-source image generation and editing model family that delivers near-closed-source performance with an order of magnitude less data.

Temporal Difference Calibration in Sequential Tasks: Application to Vision-Language-Action Models

[ICML 2026] Think Less, Act Early: Reinforced Latent Reasoning with Early Exit in Vision-Language-Action Models

[CVPR 2026] Official implementation of "ACoT-VLA: Action Chain-of-Thought for Vision-Language-Action Models"

Official project page and code repository for Enfold.

[Tech Report] Context Scaling: Scaling Properties of Text Conditioning in Visual Generation

Official repository of WCM, A World Critic Model for Vision-Language-Action Reinforcement Learning.

Implementation of RoboTTT proposed by Yunfan Jiang et al. of Stanford and Nvidia