# MolmoAct introduces action reasoning models for more explainable robot control

**URL:** <https://community.deeplearning.ai/t/molmoact-introduces-action-reasoning-models-for-more-explainable-robot-control/873827>\
**Category:** AI Discussions\
**Tags:** data-points, ai-discussions\
**Created:** [August 15, 2025, 7:53pm UTC](https://community.deeplearning.ai/t/molmoact-introduces-action-reasoning-models-for-more-explainable-robot-control/873827 "2025-08-15T19:53:34Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![Community-Team](https://sea1.discourse-cdn.com/dlai/user_avatar/community.deeplearning.ai/community-team/32/136466_2.png) [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\
**Post date:** [August 15, 2025, 7:53pm UTC](https://community.deeplearning.ai/t/molmoact-introduces-action-reasoning-models-for-more-explainable-robot-control/873827/1 "2025-08-15T19:53:34Z")

</div>

Subscribe for free access to ▶ **[Data Points!](https://www.deeplearning.ai/the-batch/tag/data-points/?utm_campaign=dp-campaign&utm_medium=announcement&utm_source=discourse)**

 ![Subscribe for free access to Data Points!](https://global.discourse-cdn.com/dlai/original/3X/6/e/6e26d5414c51b325c45786f4dac20576625a5f9e.jpeg)

Researchers from the University of Washington and Allen Institute for AI have developed MolmoAct, a family of open-source robotic foundation models that integrate perception, planning, and control through structured reasoning. The models generate three types of tokens sequentially: depth perception tokens for 3D understanding, visual reasoning traces showing planned trajectories, and action tokens for robot control. MolmoAct-7B-D achieved 70.5 percent zero-shot accuracy on SimplerEnv Visual Matching tasks, surpassing closed-source models π0 and GR00T N1 (while taking much less time to pre-train), and 86.6 percent average success on LIBERO benchmarks. This more transparent approach to model trajectories in particular addresses some limitations in current vision-language-action models, making robot decision-making more explainable and steerable through visual trajectory editing. The team released all model weights, training code, and the MolmoAct Dataset containing over 10,000 robot trajectories. ([arXiv](https://arxiv.org/abs/2508.07917))

---

<div class="post-metadata">

**Author:** ![Community-Team](https://sea1.discourse-cdn.com/dlai/user_avatar/community.deeplearning.ai/community-team/32/136466_2.png) [@Community-Team](https://community.deeplearning.ai/u/Community-Team)\
**Post date:** [August 15, 2025, 7:53pm UTC](https://community.deeplearning.ai/t/molmoact-introduces-action-reasoning-models-for-more-explainable-robot-control/873827/2 "2025-08-15T19:53:58Z")

</div>

Pinned on August 15, 2025, 7:53pm UTC