Topics / #world-action-model

World-action-model projects on GitHub

World-action models that couple environment prediction with action generation for robotics research.

As of , the most-starred World-action-model projects on GitHub are starVLA (3.8K stars), OpenWAM (1K stars) and FluxVLA (748 stars). GitHub.Locker tracks 8 World-action-model repositories with 7.1K stars combined. The fastest-growing on October 16 was video-prediction-policy-2, with +11 stars in a day.

8Projects
7.1KStars in total
+22Stars in a day
Browse in the catalog →github.com/topics/world-action-model ↗

Hot right now

Most stars gained in a day · Oct 16

+11

video-prediction-policy-2

roboterax/video-prediction-policy-2 · ★ 249

Video Prediction Policy 2 (VPP2) is a world-action model (WAM) for robot manipulation that predicts future video frames and converts them into robot actions, with strong zero-shot generalization across tasks.

+3

starVLA

starVLA/starVLA · ★ 3.8K

StarVLA is an open-source, modular research codebase for building Vision-Language-Action (VLA) models that control generalist robots, letting you mix and match backbones, action heads, and training recipes like Lego bricks.

+2

OpenWAM

OpenWAM-Official/OpenWAM · ★ 1K

OpenWAM is an open research stack for systematically building World-Action Models (WAMs) for robots, breaking the usual tightly-coupled design choices into modular, swappable components you can compare in controlled experiments.

+2

FluxVLA

FluxVLA/FluxVLA · ★ 748

FluxVLA is a full-stack engineering platform for embodied AI covering the whole loop from data to real-robot deployment for vision-language-action (VLA) models, built to lower the engineering barrier for VLA research and development.

The all-time top

8 projects · by stars

  1. #1 · starVLA/starVLA

    starVLA

    StarVLA is an open-source, modular research codebase for building Vision-Language-Action (VLA) models that control generalist robots, letting you mix and match backbones, action heads, and training recipes like Lego bricks.

    3.8K
  2. #2 · OpenWAM-Official/OpenWAM

    OpenWAM

    OpenWAM is an open research stack for systematically building World-Action Models (WAMs) for robots, breaking the usual tightly-coupled design choices into modular, swappable components you can compare in controlled experiments.

    1K
  3. #3 · FluxVLA/FluxVLA

    FluxVLA

    FluxVLA is a full-stack engineering platform for embodied AI covering the whole loop from data to real-robot deployment for vision-language-action (VLA) models, built to lower the engineering barrier for VLA research and development.

    748
  1. 4

    Zero-WAM

    434robbyant-research/Zero-WAM · Python

    Zero-WAM is a research method for zero-shot cross-task robotic manipulation: a causal video-action policy that learns in-context world-action modeling, predicting future robot observations and executable actions from either language instructions or human demonstration videos.

  2. 5

    EasyWAM

    401OpenMOSS/EasyWAM · Python

    EasyWAM is a unified research codebase for training, fine-tuning, and evaluating World Action Models (WAMs) — models that jointly predict future video and robot actions for embodied-AI and robotics research.

  3. 6

    video-prediction-policy-2

    249roboterax/video-prediction-policy-2 · Python

    Video Prediction Policy 2 (VPP2) is a world-action model (WAM) for robot manipulation that predicts future video frames and converts them into robot actions, with strong zero-shot generalization across tasks.

  4. 7

    flex-pi

    225geyan21/flex-pi · Python

    Flex-π is a 6B-parameter world-action model for robot manipulation that jointly predicts future RGB, 3D pointmaps, DINOv3 semantics, and actions from one checkpoint, so it can deploy as a VLA, a full world model, or anything in between.

  5. 8

    Embodied-AI-Start

    199ppsdk/Embodied-AI-Start

    A Chinese-language 'Embodied AI Starter Map' that organizes a learning path from robotics fundamentals (coordinates, kinematics, ROS 2) up to VLA policies, World Models, model-based RL and World Action Models (WAM), linking papers, code, benchmarks and practice guides at each step.

Often tagged together