# Reinforcing Dual-Path Reasoning in Spatial Vision Language Models

arXiv: https://arxiv.org/abs/2606.17539 · HTML: https://mlsanity.com/paper/2606.17539/

## 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 32 of its 60 references that are in the map.
- First posted on arXiv 16 Jun 2026.

## Placed in

- LLM reasoning / [Spatial and embodied reasoning](https://mlsanity.com/lr/spatial-and-embodied-reasoning/)
- Multimodal LLMs / vision-language models / [Embodied and 3D vision-language models](https://mlsanity.com/mllm/embodied-and-3d-vision-language-models/)

## Builds on (most cited first, 30 of 32)

- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond](https://mlsanity.com/paper/2308.12966/) (2023)
- [s1: Simple test-time scaling](https://mlsanity.com/paper/2501.19393/) (2025)
- [SpatialVLM: Endowing Vision-Language Models with Spatial Reasoning Capabilities](https://mlsanity.com/paper/2401.12168/) (2024)
- [Vision-R1: Incentivizing Reasoning Capability in Multimodal Large Language Models](https://mlsanity.com/paper/2503.06749/) (2025)
- [BLINK: Multimodal Large Language Models Can See but Not Perceive](https://mlsanity.com/paper/2404.12390/) (2024)
- [R1-Onevision: Advancing Generalized Multimodal Reasoning through Cross-Modal Formalization](https://mlsanity.com/paper/2503.10615/) (2025)
- [VL-Rethinker: Incentivizing Self-Reflection of Vision-Language Models with Reinforcement Learning](https://mlsanity.com/paper/2504.08837/) (2025)
- [SpatialRGPT: Grounded Spatial Reasoning in Vision Language Models](https://mlsanity.com/paper/2406.01584/) (2024)
- [MM-Eureka: Exploring the Frontiers of Multimodal Reasoning with Rule-based Reinforcement Learning](https://mlsanity.com/paper/2503.07365/) (2025)
- [R1-VL: Learning to Reason with Multimodal Large Language Models via Step-wise Group Relative Policy Optimization](https://mlsanity.com/paper/2503.12937/) (2025)
- [LLaVA-3D: A Simple yet Effective Pathway to Empowering LMMs with 3D-awareness](https://mlsanity.com/paper/2409.18125/) (2024)
- [Cosmos-Reason1: From Physical Common Sense To Embodied Reasoning](https://mlsanity.com/paper/2503.15558/) (2025)
- [Eight Methods to Evaluate Robust Unlearning in LLMs](https://mlsanity.com/paper/2402.16835/) (2024)
- [Omni3D: A Large Benchmark and Model for 3D Object Detection in the Wild](https://mlsanity.com/paper/2207.10660/) (2022)
- [SpaceR: Reinforcing MLLMs in Video Spatial Reasoning](https://mlsanity.com/paper/2504.01805/) (2025)
- [LlamaV-o1: Rethinking Step-by-step Visual Reasoning in LLMs](https://mlsanity.com/paper/2501.06186/) (2025)
- [From Flatland to Space: Teaching Vision-Language Models to Perceive and Reason in 3D](https://mlsanity.com/paper/2503.22976/) (2025)
- [InternGPT: Solving Vision-Centric Tasks by Interacting with ChatGPT Beyond Language](https://mlsanity.com/paper/2305.05662/) (2023)
- [SpatialLadder: Progressive Training for Spatial Reasoning in Vision-Language Models](https://mlsanity.com/paper/2510.08531/) (2025)
- [Grounded Reinforcement Learning for Visual Reasoning](https://mlsanity.com/paper/2505.23678/) (2025)
- [RegionGPT: Towards Region Understanding Vision Language Model](https://mlsanity.com/paper/2403.02330/) (2024)
- [SpatialReasoner: Towards Explicit and Generalizable 3D Spatial Reasoning](https://mlsanity.com/paper/2504.20024/) (2025)
- [SpatialCoT: Advancing Spatial Reasoning through Coordinate Alignment and Chain-of-Thought for Embodied Task Planning](https://mlsanity.com/paper/2501.10074/) (2025)
- [OmniVinci: Enhancing Architecture and Data for Omni-Modal Understanding LLM](https://mlsanity.com/paper/2510.15870/) (2025)
- [SSR: Enhancing Depth Perception in Vision-Language Models via Rationale-Guided Spatial Reasoning](https://mlsanity.com/paper/2505.12448/) (2025)
- [Perspective-Aware Reasoning in Vision-Language Models via Mental Imagery Simulation](https://mlsanity.com/paper/2504.17207/) (2025)
- [Cubify Anything: Scaling Indoor 3D Object Detection](https://mlsanity.com/paper/2412.04458/) (2024)
- [Sparkle: Mastering Basic Spatial Capabilities in Vision Language Models Elicits Generalization to Spatial Reasoning](https://mlsanity.com/paper/2410.16162/) (2024)
- [3D Aware Region Prompted Vision Language Model](https://mlsanity.com/paper/2509.13317/) (2025)

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

- [From Sight to Foresight: Predictive Spatial Reasoning in Vision-Language Models](https://mlsanity.com/paper/2610.04139/) (2026)

## Abstract (arXiv)

Spatial VLMs have made substantial progress in geometric perception, yet complex spatial reasoning requiring multi-step inference over depth, distance, and scene relations remains challenging. Moreover, different spatial queries call for fundamentally different strategies: some are best addressed through purely linguistic, step-by-step deduction, while others require explicit 3D grounding before quantitative inference. We present Dual-Path Spatial Reasoning via Reinforcement Learning for Spatial VLMs (SR-REAL), a unified framework that equips a spatial VLM with two complementary reasoning paths: Language-Only Reasoning (LOR), which performs step-by-step linguistic deduction, and Detect-Then-Reason (DTR), which detects 3D geometric cues (e.g., centers or bounding boxes) via region tokens before explicit geometric inference. SR-REAL begins with a cold-start supervised fine-tuning stage that constructs LOR and DTR chain-of-thought supervision and exposes a region-to-3D interface, followed by RL that optimizes the policy model with accuracy and format rewards; for DTR, a discrete center-based detection reward further refines geometric alignment. Across diverse spatial benchmarks, SR-REAL significantly outperforms spatial VLM baselines: (i) a single RL-trained model supports both reasoning paths, with DTR excelling in region-aware tasks through precise 3D localization and LOR enhancing general spatial reasoning; (ii) jointly training both paths fosters mutual reinforcement; (iii) high-quality, blended cold-start data is crucial for stable RL optimization; and (iv) the model generalizes across datasets and domains without per-task tuning, demonstrating positive transfer between LOR and DTR.