JoVox jointly models voxel hierarchy with discrete flow matching
Authors report 8x training throughput, 1/15 forward FLOPs, and 11% lower Chamfer distance than TRELLIS.2.
Stories per day · 30 days
2 RSS
Relatedcoding-agents3scene-generation3benchmarks2blender23dcodeverse13d-editing1agents1animation1claude1claude-opus1code-editing1computer-vision1critique-models1dataset1datasets1diffusion-transformer1diffusion-transformers1dinov21
31 stories, newest first · one card per story
Stories tagged #3d-generation
CCORE GLB-3D-Coder 4B model trending on Hugging Face
4B model converts text prompts to editable 3D assets exportable as GLB files for Godot or Three.js.
1
SeMo3D unifies 3D generation editing and completion
SeMo3D uses selective modeling to unify 3D tasks with region-selective denoising and Omni3DiT architecture.
DyCraft breaks long-horizon 3D animation into storyboard steps for coding agents
On 24 multi-shot tasks, human raters preferred DyCraft over a native GPT-6 agent in Codex and prior baselines.
3DCodeEdit benchmark: models average 37% on 3D code edits
Reviewer-model feedback on renders helps more than extra reasoning or an agent harness, and can lift a 27B model past a 2.8T model's single-shot rate.
Paper generates 3D scenes in implicit latents from 2D models
Diffusion transformers model scenes in a 3D latent from DINOv2, SigLIP2, and DA3 instead of redundant 2D views.
Scene-Designer learns 3D scene design and critique with critic-guided GRPO
Two new datasets train design and critique, and test-time critic guidance further improves generation.
3DCodeVerse: Exploring 3D Worlds as Code
3DCodeVerse pairs a 2M executable-code dataset with a harness that beats Claude Design in tests.
SUPRA-TEX: 3D Generation with Super-Aligned Texture
SUPRA-Tex textures 3D meshes from one image using aligned PBR latents and reports commercial-level fidelity.
3DCodeVerse: Exploring 3D Worlds as Code
3DCodeVerse pairs a 2M executable-3D dataset and verification harness that wins 90% against Claude Design.
BlenderWorld benchmarks frontier models on real-world 3D creation
GPT-6-Astra scored 79.9 overall, ahead of Opus-5 at 73.0 and GPT-5.6-Sol at 67.7.
Towards Physically Grounded Articulated 3D Assets: A Multi-Agent Pipeline and Benchmark
Artisan’s multi-agent pipeline adds physics feedback so generated articulated 3D assets collide less and stay more stable.
SpaceFlow: Locally Controllable 3D Generation
SpaceFlow adds local geometric and appearance control to training-free text-to-3D generation.
SpaceFlow: Locally Controllable 3D Generation
SpaceFlow enables locally controllable, training-free 3D generation from text and geometric primitives.
Tetris3D: 3D Scene Generation With Objects That Fit Together
Tetris3D reconstructs single-image 3D scenes so interacting objects remain geometrically and physically coherent.
Building Rome from a Single Image
A redesigned object generator builds complete indoor and outdoor 3D meshes from one image, beating prior baselines.
MC-Sparse: Deconstructing and Closing the Dense-Sparse Attention Gap in Diffusion Transformers
Training-free MC-Sparse speeds diffusion-transformer denoising up to 2.32× with negligible quality loss.
MC-Sparse: Deconstructing and Closing the Dense-Sparse Attention Gap in Diffusion Transformers
MC-Sparse narrows the dense-sparse attention gap, speeding diffusion denoising up to 2.32× with little quality loss.
DistScene: Object-to-Scene Distillation for 3D Scene Generation
DistScene generates compositional 3D scenes from one image by jointly modeling environments and objects.
LMBuild: Evaluating LLM Agents for Generating Buildable and Functional Structures
LMBuild benchmarks LLM agents on generating structures that are buildable, functional, and physically realizable.
SILSA: Sliding-Window Slice Latents for Topology-Preserving High-Resolution 3D Generation
SILSA generates high-resolution 3D shapes from sliding-window slice latents, using far fewer tokens and improving topology.
Octrees as an Explicit 3D Language
OctLLM feeds 3D shapes to an LLM as sparse octree tokens via frozen-backbone trainable branches, cutting image-to-3D FID 17.4% versus ShapeLLM-Omni.
SILSA: Sliding-Window Slice Latents for Topology-Preserving High-Resolution 3D Generation
SILSA generates high-resolution 3D shapes from compact slice latents while preserving topology and cutting token cost.
GeoVerse: World-Consistent Novel View Synthesis in Geometric Latent Space
GeoVerse synthesizes world-consistent novel views inside a geometric latent space using video appearance priors.