# 3DThinkVLA: Endowing Vision-Language-Action Models with Latent 3D Priors via 3D-Thinking-Guided Co-training

arXiv: https://arxiv.org/abs/2606.04436 · HTML: https://mlsanity.com/paper/2606.04436/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Spatial and embodied reasoning, one of 17 directions in LLM reasoning.
- Builds on 37 of its 46 references that are in the map.
- First posted on arXiv 3 Jun 2026.

## Placed in

- LLM reasoning / [Spatial and embodied reasoning](https://mlsanity.com/lr/spatial-and-embodied-reasoning/)
- Vision-language-action models / [3D and spatial-aware VLA](https://mlsanity.com/vla/3d-and-spatial-aware-vla/)

## Builds on (most cited first, 30 of 37)

- [RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control](https://mlsanity.com/paper/2307.15818/) (2023)
- [LLaVA-OneVision: Easy Visual Task Transfer](https://mlsanity.com/paper/2408.03326/) (2024)
- [RT-1: Robotics Transformer for Real-World Control at Scale](https://mlsanity.com/paper/2212.06817/) (2022)
- [VGGT: Visual Geometry Grounded Transformer](https://mlsanity.com/paper/2503.11651/) (2025)
- [LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learning](https://mlsanity.com/paper/2306.03310/) (2023)
- [FAST: Efficient Action Tokenization for Vision-Language-Action Models](https://mlsanity.com/paper/2501.09747/) (2025)
- [Point Transformer V3: Simpler, Faster, Stronger](https://mlsanity.com/paper/2312.10035/) (2023)
- [SpatialVLA: Exploring Spatial Representations for Visual-Language-Action Model](https://mlsanity.com/paper/2501.15830/) (2025)
- [Evaluating Real-World Robot Manipulation Policies in Simulation](https://mlsanity.com/paper/2405.05941/) (2024)
- [CoT-VLA: Visual Chain-of-Thought Reasoning for Vision-Language-Action Models](https://mlsanity.com/paper/2503.22020/) (2025)
- [3D-VLA: A 3D Vision-Language-Action Generative World Model](https://mlsanity.com/paper/2403.09631/) (2024)
- [LIBERO-Plus: In-depth Robustness Analysis of Vision-Language-Action Models](https://mlsanity.com/paper/2510.13626/) (2025)
- [Spatial Forcing: Implicit Spatial Representation Alignment for Vision-language-action Model](https://mlsanity.com/paper/2510.12276/) (2025)
- [NORA: A Small Open-Sourced Generalist Vision Language Action Model for Embodied Tasks](https://mlsanity.com/paper/2504.19854/) (2025)
- [PointVLA: Injecting the 3D World into Vision-Language-Action Models](https://mlsanity.com/paper/2503.07511/) (2025)
- [ABot-M0: VLA Foundation Model for Robotic Manipulation with Action Manifold Learning](https://mlsanity.com/paper/2602.11236/) (2026)
- [Self-Improving Vision-Language-Action Models with Data Generation via Residual RL](https://mlsanity.com/paper/2511.00091/) (2025)
- [GeoVLA: Empowering 3D Representations in Vision-Language-Action Models](https://mlsanity.com/paper/2508.09071/) (2025)
- [4D-VLA: Spatiotemporal Vision-Language-Action Pretraining with Cross-Scene Calibration](https://mlsanity.com/paper/2506.22242/) (2025)
- [Lift3D Foundation Policy: Lifting 2D Large-Scale Pretrained Models for Robust 3D Robotic Manipulation](https://mlsanity.com/paper/2411.18623/) (2024)
- [DepthVLA: Enhancing Vision-Language-Action Models with Depth-Aware Spatial Reasoning](https://mlsanity.com/paper/2510.13375/) (2025)
- [VLM4VLA: Revisiting Vision-Language-Models in Vision-Language-Action Models](https://mlsanity.com/paper/2601.03309/) (2026)
- [FP3: A 3D Foundation Policy for Robotic Manipulation](https://mlsanity.com/paper/2503.08950/) (2025)
- [3D CAVLA: Leveraging Depth and 3D Context to Generalize Vision Language Action Models for Unseen Tasks](https://mlsanity.com/paper/2505.05800/) (2025)
- [StarVLA-α: Reducing Complexity in Vision-Language-Action Systems](https://mlsanity.com/paper/2604.11757/) (2026)
- [From Spatial to Actions: Grounding Vision-Language-Action Model in Spatial Foundation Priors](https://mlsanity.com/paper/2510.17439/) (2025)
- [Pre-training Auto-regressive Robotic Models with 4D Representations](https://mlsanity.com/paper/2502.13142/) (2025)
- [GeoAware-VLA: Implicit Geometry Aware Vision-Language-Action Model](https://mlsanity.com/paper/2509.14117/) (2025)
- [QDepth-VLA: Quantized Depth Prediction as Auxiliary Supervision for Vision-Language-Action Models](https://mlsanity.com/paper/2510.14836/) (2025)
- [DualCoT-VLA: Visual-Linguistic Chain of Thought via Parallel Reasoning for Vision-Language-Action Models](https://mlsanity.com/paper/2603.22280/) (2026)

## Built on by (most cited first, 1 of 1)

- [FOCAL-VLA: Subtask-Guided Geometry Distillation and Implicit World Modeling for Vision-Language-Action Models](https://mlsanity.com/paper/2609.21228/) (2026)

## Abstract (arXiv)

We propose a 3D-thinking-guided co-training framework that enables vision-language-action (VLA) models to perform 3D spatial reasoning implicitly during action prediction. Our core insight is that 3D geometry perception and 3D spatial reasoning are distinct capabilities that can be disentangled and injected at different feature hierarchies. During training, three tightly coupled components work in concert primarily within the latent space: (1) To gain geometric priors, a latent 3D geometry perception module aligns intermediate visual features with a 3D foundation model, acquiring low-level geometric cues without architectural modifications to the VLM backbone. (2) Complementing this, an online 3D reasoning distillation module mitigates the prompt-induced reasoning gap via a shared reasoning anchor token. During 3D VLM co-training, this anchor is emitted as the first output token to robustly encode spatial priors. During VLA training, it serves as an input token inserted between the task and action instructions, transferring high-level spatial thinking from explicit teacher reasoning prompts to student action prompts without chain-of-thought text generation. (3) These disentangled geometric and reasoning features are then united by a spatially augmented action integration, which jointly injects them into the action-query tokens as hierarchical spatial conditions to prevent action shortcuts. At deployment, our method retains only its lightweight adapters to perform implicit 3D reasoning, discarding the 3D foundation model and the teacher branch used for supervision. Consequently, it operates purely on 2D images without 3D sensors, external models, or explicit text generation while preventing catastrophic forgetting of the pretrained VLM, achieving state-of-the-art performance on LIBERO, LIBERO-PLUS, SimplerEnv, and real-world manipulation tasks.