DeepThinkVLA: Enhancing Reasoning Capability of Vision-Language-Action Models

• URL: http://arxiv.org/abs/2511.15669v1
• Date: Fri, 31 Oct 2025 05:26:16 GMT
• Title: DeepThinkVLA: Enhancing Reasoning Capability of Vision-Language-Action Models
• Authors: Cheng Yin, Yankai Lin, Wang Xu, Sikyuen Tam, Xiangrui Zeng, Zhiyuan Liu, Zhouping Yin,
• Abstract summary: DeepThinkVLA is a new architecture for Vision-Language-Action models.<n>It generates sequential CoT with causal attention and switches to bidirectional attention for fast decoding of action vectors.<n>It achieves a 97.0% success rate on the LIBERO benchmark.
• Score: 51.76664843721462
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Enabling Vision-Language-Action (VLA) models to "think before acting" via Chain-of-Thought (CoT) is a promising path to overcoming the data-hungry nature of end-to-end robot policies. However, progress is stalled by a fundamental conflict: existing models use a single autoregressive decoder for both sequential CoT reasoning and high-dimensional, parallelizable robot actions. This architectural mismatch degrades motor control and fails to forge a strong causal link between thought and action. We introduce DeepThinkVLA, which resolves this conflict through a tightly integrated architecture and training strategy. Architecturally, our hybrid-attention decoder generates sequential CoT with causal attention and then switches to bidirectional attention for fast, parallel decoding of action vectors. This design is complemented by a two-stage training pipeline: we first use Supervised Fine-Tuning (SFT) to teach the model foundational reasoning, then apply Reinforcement Learning (RL) with task-success rewards to causally align the full reasoning-action sequence with desired outcomes. This synergy leads to state-of-the-art performance, achieving a 97.0% success rate on the LIBERO benchmark. Our ablations confirm the design's effectiveness: the hybrid architecture alone outperforms standard decoders by 15.5%, and the final RL stage provides a crucial 2% boost to secure top performance.

Related papers

• ActionCodec: What Makes for Good Action Tokenizers [106.78093973045526]
  Vision-Language-Action (VLA) models have demonstrated superior instruction-following and training efficiency.<n>Central to this paradigm is action tokenization, yet its design has primarily focused on reconstruction fidelity.<n>We introduce textbfActionCodec, a high-performance action tokenizer that significantly enhances both training efficiency and VLA performance.
  arXiv  Detail & Related papers  (2026-02-17T07:07:15Z)
• FASTer: Toward Efficient Autoregressive Vision Language Action Modeling via Neural Action Tokenization [61.10456021136654]
  We introduce FASTer, a unified framework for efficient and general robot learning.<n>FASTerVQ encodes action chunks as single-channel images, capturing global-temporal dependencies while maintaining a high compression ratio.<n>FASTerVLA builds on this tokenizer with block-wise autoregressive decoding and a lightweight action expert, achieving both faster inference and higher task performance.
  arXiv  Detail & Related papers  (2025-12-04T16:21:38Z)
• DualVLA: Building a Generalizable Embodied Agent via Partial Decoupling of Reasoning and Action [62.70893433854428]
  We propose DualVLA, which enhances action performance through carefully designed post-training while still preserving reasoning capability.<n>Experiments show that DualVLA achieves an average success rate of 61.0 in SimplerEnv and an average score of 65.4 across eight competitive multimodal benchmarks.
  arXiv  Detail & Related papers  (2025-11-27T06:03:53Z)
• VLA-R1: Enhancing Reasoning in Vision-Language-Action Models [35.264042764326895]
  Vision-Language-Action (VLA) models aim to unify perception, language understanding, and action generation.<n>Current VLA models often lack explicit step-by-step reasoning.<n>We present VLA-R1, a reasoning-enhanced VLA that integrates Reinforcement Learning from Verifiable Rewards.
  arXiv  Detail & Related papers  (2025-10-02T02:54:03Z)
• How LLMs Learn to Reason: A Complex Network Perspective [14.638878448692493]
  Training large language models with Reinforcement Learning from Verifiable Rewards exhibits a set of puzzling behaviors.<n>We propose that these seemingly disparate phenomena can be explained using a single unifying theory.<n>Our work provides a new physical intuition for engineering the emergent reasoning capabilities of future AI systems.
  arXiv  Detail & Related papers  (2025-09-28T04:10:37Z)
• Efficient Generative Model Training via Embedded Representation Warmup [12.485320863366411]
  Generative models face a fundamental challenge: they must simultaneously learn high-level semantic concepts and low-level synthesis details.<n>We propose Embedded Representation Warmup, a principled two-phase training framework.<n>Our framework achieves a 11.5$times$ speedup in 350 epochs to reach FID=1.41 compared to single-phase methods like REPA.
  arXiv  Detail & Related papers  (2025-04-14T12:43:17Z)
• End-to-End Meta-Bayesian Optimisation with Transformer Neural Processes [52.818579746354665]
  This paper proposes the first end-to-end differentiable meta-BO framework that generalises neural processes to learn acquisition functions via transformer architectures. We enable this end-to-end framework with reinforcement learning (RL) to tackle the lack of labelled acquisition data.
  arXiv  Detail & Related papers  (2023-05-25T10:58:46Z)
• Building Robust Ensembles via Margin Boosting [98.56381714748096]
  In adversarial robustness, a single model does not usually have enough power to defend against all possible adversarial attacks. We develop an algorithm for learning an ensemble with maximum margin. We show that our algorithm not only outperforms existing ensembling techniques, but also large models trained in an end-to-end fashion.
  arXiv  Detail & Related papers  (2022-06-07T14:55:58Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.