mlsanityData to 7 Oct 2026

VoxelCodeBench: Benchmarking 3D World Modeling Through Code Generation

arXiv 2604.02580 ↗First posted 2 Apr 2026cs.LGcited by 7

On the map: World models → Video-generation world models and simulators

also in 1 more
Abstract (arXiv)

Evaluating code generation models for 3D spatial reasoning requires executing generated code in realistic environments and assessing outputs beyond surface-level correctness. We introduce a platform VoxelCode, for analyzing code generation capabilities for 3D understanding and environment creation. Our platform integrates natural language task specification, API-driven code execution in Unreal Engine, and a unified evaluation pipeline supporting both automated metrics and human assessment. To demonstrate its utility, we construct VoxelCodeBench, a benchmark of voxel manipulation tasks spanning three reasoning dimensions: symbolic interpretation, geometric construction, and artistic composition. Evaluating leading code generation models, we find that producing executable code is far easier than producing spatially correct outputs, with geometric construction and multi-object composition proving particularly challenging. By open-sourcing our platform and benchmark, we provide the community with extensible infrastructure for developing new 3D code generation benchmarks and probing spatial reasoning in future models.

Key facts. Cited by 7 papers in the map (data through 7 Oct 2026). Placed in Video-generation world models and simulators, one of 11 directions in World models. Builds on 29 of its 41 references that are in the map. First posted on arXiv 2 Apr 2026.

Builds on

29 of 41 references.

Closest ancestors · 1

Same direction, newest first

  1. 2022DreamFusion: Text-to-3D using 2D Diffusion29 Sep

Other references · 28

  1. 2021Evaluating Large Language Models Trained on Codecited by 8,053
  2. 2021Program Synthesis with Large Language Modelscited by 3,509
  3. 2023Voyager: An Open-Ended Embodied Agent with Large Language Modelscited by 2,081
  4. 2022Competition-Level Code Generation with AlphaCodecited by 1,427
  5. 2022Magic3D: High-Resolution Text-to-3D Content Creationcited by 1,285
  6. 2023ProlificDreamer: High-Fidelity and Diverse Text-to-3D Generation with Variational Score Distillationcited by 1,207
  7. 2022Code as Policies: Language Model Programs for Embodied Controlcited by 1,172
  8. 2019Habitat: A Platform for Embodied AI Researchcited by 1,163
  9. 2021Measuring Coding Challenge Competence With APPScited by 879
  10. 2023DreamGaussian: Generative Gaussian Splatting for Efficient 3D Content Creationcited by 837
  11. 2019ALFRED: A Benchmark for Interpreting Grounded Instructions for Everyday Taskscited by 773
  12. 2023The Dawn of LMMs: Preliminary Explorations with GPT-4V(ision)cited by 677
16 more
  1. 2022ProgPrompt: Generating Situated Robot Task Plans using Large Language Modelscited by 598
  2. 20163D-R2N2: A Unified Approach for Single and Multi-view 3D Object Reconstructioncited by 587
  3. 2023API-Bank: A Comprehensive Benchmark for Tool-Augmented LLMscited by 574
  4. 2022MineDojo: Building Open-Ended Embodied Agents with Internet-Scale Knowledgecited by 469
  5. 2022DS-1000: A Natural and Reliable Benchmark for Data Science Code Generationcited by 442
  6. 2016OctNet: Learning Deep 3D Representations at High Resolutionscited by 340
  7. 2023OmniObject3D: Large-Vocabulary 3D Object Dataset for Realistic Perception, Reconstruction and Generationcited by 324
  8. 2023GPT4RoI: Instruction Tuning Large Language Model on Region-of-Interestcited by 300
  9. 2023CodeGeeX: A Pre-Trained Model for Code Generation with Multilingual Benchmarking on HumanEval-Xcited by 294
  10. 2022Visual Spatial Reasoningcited by 288
  11. 2024SceneCraft: An LLM Agent for Synthesizing 3D Scene as Blender Codecited by 125
  12. 2021SpartQA: : A Textual Question Answering Benchmark for Spatial Reasoningcited by 75
  13. 20233D generation on ImageNetcited by 47
  14. 2024BlenderLLM: Training Large Language Models for Computer-Aided Design with Self-improvementcited by 26
  15. 2025MeshCoder: LLM-Powered Structured Mesh Code Generation from Point Cloudscited by 15
  16. 2025PrimitiveAnything: Human-Crafted 3D Primitive Assembly Generation with Auto-Regressive Transformercited by 12

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 7.

7 papers.

  1. 20263DCodeBench: Benchmarking Agentic Procedural 3D Modeling Via Codecited by 9
  2. 2026P3D-Bench: Benchmarking MLLMs for Parametric 3D Generation and Structural Reasoningcited by 8
  3. 2026Code as Worlds: Agentic Discovery of Executable World Representations for Physical Reasoningcited by 6
  4. 2026SimWorlds: A Multi-Agent System for Dynamic 3D Scene Creationcited by 4
  5. 2026Sample-Efficient Post-Training for LEGO Spatial-Physics Reasoningcited by 1
  6. 2026Agentic Game Development as a Verifiable Trajectory Data Engine for Scaling World Modelscited by 1
  7. 20263DHarnessBench: Probing Agentic 3D-to-Code Capabilities of Frontier Vision-Language Modelscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Code4Scene: Benchmarking Coding Agents for Constructing and Editing 3D Scenes
  2. 2026WorldBench: Evaluating LLMs on Three.js Voxel World Generation
  3. 2026How Far Are Vision-Language Models from Constructing the Real World? A Benchmark for Physical Generative Reasoning
  4. 20243DBench: A Scalable 3D Benchmark and Instruction-Tuning Dataset
  5. 2025GamiBench: Evaluating Spatial Reasoning and 2D-to-3D Planning Capabilities of MLLMs with Origami Folding Tasks
  6. 2026Flame3D: Zero-shot Compositional Reasoning of 3D Scenes with Agentic Language Models
  7. 20243DSRBench: A Comprehensive 3D Spatial Reasoning Benchmark
  8. 2026GraFT: A Training-Free Framework for Spatial Reasoning in Multimodal Large Language Models via 3D Scene Graphs