# FOCAL-VLA: Subtask-Guided Geometry Distillation and Implicit World Modeling for Vision-Language-Action Models

arXiv: https://arxiv.org/abs/2609.21228 · HTML: https://mlsanity.com/paper/2609.21228/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Embodied and robot world models, one of 11 directions in World models.
- Builds on 17 of its 46 references that are in the map.
- First posted on arXiv 18 Sep 2026.

## Placed in

- World models / [Embodied and robot world models](https://mlsanity.com/wm/embodied-and-robot-world-models/)
- Multimodal LLMs / vision-language models / [Embodied and 3D vision-language models](https://mlsanity.com/mllm/embodied-and-3d-vision-language-models/)
- Vision-language-action models / [3D and spatial-aware VLA](https://mlsanity.com/vla/3d-and-spatial-aware-vla/)

## Builds on (most cited first, 17 of 17)

- [Qwen3-VL Technical Report](https://mlsanity.com/paper/2511.21631/) (2025)
- [SpatialVLA: Exploring Spatial Representations for Visual-Language-Action Model](https://mlsanity.com/paper/2501.15830/) (2025)
- [π0.7: a Steerable Generalist Robotic Foundation Model with Emergent Capabilities](https://mlsanity.com/paper/2604.15483/) (2026)
- [A Pragmatic VLA Foundation Model](https://mlsanity.com/paper/2601.18692/) (2026)
- [GeoVLA: Empowering 3D Representations in Vision-Language-Action Models](https://mlsanity.com/paper/2508.09071/) (2025)
- [Being-H0.5: Scaling Human-Centric Robot Learning for Cross-Embodiment Generalization](https://mlsanity.com/paper/2601.12993/) (2026)
- [HiF-VLA: Hindsight, Insight and Foresight through Motion Representation for Vision-Language-Action Models](https://mlsanity.com/paper/2512.09928/) (2025)
- [Track4World: Feedforward World-centric Dense 3D Tracking of All Pixels](https://mlsanity.com/paper/2603.02573/) (2026)
- [GLaD: Geometric Latent Distillation for Vision-Language-Action Models](https://mlsanity.com/paper/2512.09619/) (2025)
- [ROCKET: Residual-Oriented Multi-Layer Alignment for Spatially-Aware Vision-Language-Action Models](https://mlsanity.com/paper/2602.17951/) (2026)
- [LaMP: Learning Vision-Language-Action Policy with 3D Scene Flow as Latent Motion Prior](https://mlsanity.com/paper/2603.25399/) (2026)
- [Pri4R: Learning World Dynamics for Vision-Language-Action Models with Privileged 4D Representation](https://mlsanity.com/paper/2603.01549/) (2026)
- [Track4Action: Distilling World-Centric 3D Tracker into Vision-Language-Action Policies](https://mlsanity.com/paper/2608.03727/) (2026)
- [Spatially Aware World Action Model via Geometric Latent Diffusion](https://mlsanity.com/paper/2609.02531/) (2026)
- [GaussianDream: A Feed-Forward 3D Gaussian World Model for Robotic Manipulation](https://mlsanity.com/paper/2605.20752/) (2026)
- [3DThinkVLA: Endowing Vision-Language-Action Models with Latent 3D Priors via 3D-Thinking-Guided Co-training](https://mlsanity.com/paper/2606.04436/) (2026)
- [See like a Robot: Robot-Centric Pointmaps for VLA Models](https://mlsanity.com/paper/2607.11498/) (2026)

## Abstract (arXiv)

Vision-language-action (VLA) models built on pretrained vision-language models have demonstrated strong performance across diverse robotic manipulation tasks. However, VLA models that directly map current 2D observations to actions often lack sufficient spatial and temporal understanding, limiting their performance in precise and long-horizon manipulation. Recent methods enhance VLA models through geometric supervision and future-state prediction across the entire scene. However, these methods can suffer from redundant scene information, distracting the model from learning the geometry and dynamics relevant to the current interaction. To address this issue, we propose FOCAL-VLA, a framework that combines subtask-guided geometry distillation with implicit world modeling to learn representations of current spatial structure and future interaction dynamics. To focus geometric learning on the current subtask, we transfer geometric knowledge from VGGT to the VLA model by aligning geometry latents with features from subtask-relevant image regions. To capture the future 3D evolution of the current interaction, we incorporate implicit world modeling using Track4World features from current and future demonstration frames. The two complementary representations jointly guide action generation without running VGGT or Track4World at inference time. Experiments show that FOCAL-VLA outperforms baselines on both simulation benchmarks and real-world manipulation tasks. Project website: https://zhiyuan-gao.github.io/FOCAL-VLA/.