SR-REAL equips spatial VLMs with dual LOR and DTR reasoning paths trained via RL, achieving better benchmark performance through mutual reinforcement and generalization without per-task tuning.
hub
Spatialreasoner: Towards explicit and generalizable 3d spatial reasoning
15 Pith papers cite this work. Polarity classification is still indexing.
15
Pith papers citing it
abstract
Despite recent advances on multi-modal models, 3D spatial reasoning remains a challenging task for state-of-the-art open-source and proprietary models. Recent studies explore data-driven approaches and achieve enhanced spatial reasoning performance by fine-tuning models on 3D-related visual question-answering data. However, these methods typically perform spatial reasoning in an implicit manner and often fail on questions that are trivial to humans, even with long chain-of-thought reasoning. In this work, we introduce SpatialReasoner, a novel large vision-language model (LVLM) that addresses 3D spatial reasoning with explicit 3D representations shared between multiple stages--3D perception, computation, and reasoning. Explicit 3D representations provide a coherent interface that supports advanced 3D spatial reasoning and improves the generalization ability to novel question types. Furthermore, by analyzing the explicit 3D representations in multi-step reasoning traces of SpatialReasoner, we study the factual errors and identify key shortcomings of current LVLMs. Results show that our SpatialReasoner achieves improved performance on a variety of spatial reasoning benchmarks, outperforming Gemini 2.0 by 9.2% on 3DSRBench, and generalizes better when evaluating on novel 3D spatial reasoning questions. Our study bridges the 3D parsing capabilities of prior visual foundation models with the powerful reasoning abilities of large language models, opening new directions for 3D spatial reasoning.
hub tools
citation-role summary
background 2
citation-polarity summary
years
2026 15roles
background 2polarities
background 2representative citing papers
A new consistency-verifier RL framework with OT-GRPO raises spatial reasoning accuracy in LRMs to near supervised levels using only internal geometric and semantic checks.
ReRe boosts open-source MLLMs on spatial reasoning benchmarks VSI-Bench and STI-Bench to rival proprietary SOTA by using a two-phase Reason then Re-reason process with Geometry-to-Video novel view synthesis.
A controlled study finds that block-wise state-space recurrence outperforms other memory designs for open-domain scene return in action-conditioned video models, and that standard replay metrics do not adequately measure memory quality.
LRMs show a large production-evaluation gap on the VAIR dataset with valid answers but invalid reasoning, driven by answer confirmation bias as evidenced by CoT analysis, linear probes, and causal patching.
VLMs fail to ground numerical values in spatial perception on new bidirectional tasks, relying on shallow cues instead of coordinate-aware representations.
Pretraining a vision-language model to output discrete 3D pose tokens on large non-robotic data, before training a robot action head, improves downstream manipulation success and data efficiency.
XEmbodied achieves SOTA on 18 embodied VQA benchmarks by fusing 3D geometric tokens and distilled physical cues into a 30B VLM with progressive curriculum training.
OmniView-Space framework with MPSM, tool-guided reasoning, and distillation achieves SOTA on spatial reasoning benchmarks for MLLMs while reducing external geometry dependencies.
SpaCE derives four theoretical results on spatial capacity, sample complexity, generalization, and bias-variance trade-offs for multi-frame MLLM reasoning, validated on MultiSPA, CA-VQA, and SpatialRGPT.
AlloSpatial adds structured allocentric priors and a harness for tool-use and arbitration to improve spatial reasoning in foundation models, with 5-18% gains on VSI-Bench and MindCube in training-free settings and further gains after RL internalization.
PASR performs pose-aware analysis-by-synthesis by aligning 3D projections with DINOv3 patch features, outperforming prior methods on clean and occluded retrieval while also handling pose estimation and classification.
A heterogeneous multi-agent framework with Bayesian test-time orchestration improves spatial reasoning across four benchmarks without parameter updates.
XEmbodied is a foundation model that integrates 3D geometric and physical signals into VLMs using a 3D Adapter and Efficient Image-Embodied Adapter, plus progressive curriculum and RL post-training, to improve spatial reasoning and embodied performance on 18 benchmarks.
OpenWorldLib defines world models as perception-centered systems with interaction and long-term memory, and provides a modular inference codebase unifying interactive video, 3D, reasoning, and VLA tasks.
citing papers explorer
-
Reinforcing Dual-Path Reasoning in Spatial Vision Language Models
SR-REAL equips spatial VLMs with dual LOR and DTR reasoning paths trained via RL, achieving better benchmark performance through mutual reinforcement and generalization without per-task tuning.
-
The Art of Interrogation: Consistency Amplifies Factuality in Spatial Reasoning
A new consistency-verifier RL framework with OT-GRPO raises spatial reasoning accuracy in LRMs to near supervised levels using only internal geometric and semantic checks.
-
Reason, Then Re-reason: Cross-view Revisiting Improves Spatial Reasoning
ReRe boosts open-source MLLMs on spatial reasoning benchmarks VSI-Bench and STI-Bench to rival proprietary SOTA by using a two-phase Reason then Re-reason process with Geometry-to-Video novel view synthesis.
-
Echo-Memory: A Controlled Study of Memory in Action World Models
A controlled study finds that block-wise state-space recurrence outperforms other memory designs for open-domain scene return in action-conditioned video models, and that standard replay metrics do not adequately measure memory quality.
-
An Enigma of Artificial Reason: Investigating the Production-Evaluation Gap in Large Reasoning Models
LRMs show a large production-evaluation gap on the VAIR dataset with valid answers but invalid reasoning, driven by answer confirmation bias as evidenced by CoT analysis, linear probes, and causal patching.
-
SPACENUM: Revisiting Spatial Numerical Understanding in VLMs
VLMs fail to ground numerical values in spatial perception on new bidirectional tasks, relying on shallow cues instead of coordinate-aware representations.
-
PoseVLA: Universal Pose Pretraining for Generalizable Vision-Language-Action Policies
Pretraining a vision-language model to output discrete 3D pose tokens on large non-robotic data, before training a robot action head, improves downstream manipulation success and data efficiency.
-
Steadily moving semi-infinite fracture in plane poroelasticity
XEmbodied achieves SOTA on 18 embodied VQA benchmarks by fusing 3D geometric tokens and distilled physical cues into a 30B VLM with progressive curriculum training.
-
OmniView-Space: Reinforcing Spatial Reasoning via Multi-Perspective Spatial Mapping
OmniView-Space framework with MPSM, tool-guided reasoning, and distillation achieves SOTA on spatial reasoning benchmarks for MLLMs while reducing external geometry dependencies.
-
SpaCE: Rethinking Spatial Capacity and Generalization in Multi-Frame Multimodal Large Language Models
SpaCE derives four theoretical results on spatial capacity, sample complexity, generalization, and bias-variance trade-offs for multi-frame MLLM reasoning, validated on MultiSPA, CA-VQA, and SpatialRGPT.
-
AlloSpatial: Agentic Harness Framework for Spatial Reasoning in Foundation Models
AlloSpatial adds structured allocentric priors and a harness for tool-use and arbitration to improve spatial reasoning in foundation models, with 5-18% gains on VSI-Bench and MindCube in training-free settings and further gains after RL internalization.
-
PASR: Pose-Aware 3D Shape Retrieval from Occluded Single Views
PASR performs pose-aware analysis-by-synthesis by aligning 3D projections with DINOv3 patch features, outperforming prior methods on clean and occluded retrieval while also handling pose estimation and classification.
-
SpatiO: Adaptive Test-Time Orchestration of Vision-Language Agents for Spatial Reasoning
A heterogeneous multi-agent framework with Bayesian test-time orchestration improves spatial reasoning across four benchmarks without parameter updates.
-
XEmbodied: A Foundation Model with Enhanced Geometric and Physical Cues for Large-Scale Embodied Environments
XEmbodied is a foundation model that integrates 3D geometric and physical signals into VLMs using a 3D Adapter and Efficient Image-Embodied Adapter, plus progressive curriculum and RL post-training, to improve spatial reasoning and embodied performance on 18 benchmarks.
-
OpenWorldLib: A Unified Codebase and Definition of Advanced World Models
OpenWorldLib defines world models as perception-centered systems with interaction and long-term memory, and provides a modular inference codebase unifying interactive video, 3D, reasoning, and VLA tasks.