Quanta AI Labs

3D & Spatial

3D generation, reconstruction, PBR materials, depth estimation, and spatial computing. 221 releases.

Language ModelsVision & ImageVideo & AnimationAudio & Speech3D & SpatialAgents & AutomationDeveloper ToolsResearch & Safety

221 releases

  • PhysAg3D by Peking University, Guangdong Laboratory of Artificial Intelligence and Digital Economy (SZ), Huawei and Wuhan University

    PhysAg3D

    PhysAg3D is a research framework that turns a single image into a textured 3D asset that can move or deform plausibly.

  • Mundo: Customized World Model

    Mundo is a customised world model that builds an interactive, editable 3D environment from a small set of context…

  • Idol by Speridlabs (Alejandro Cañada, Chema Garabito)

    Idol

    Idol is a 0.66B-parameter model from Speridlabs that turns a single image (a photograph, a render or an Iris-3B…

  • GPT-6.1 Terra Blender test by J A Z I I (@notjazii)

    GPT-6.1 Terra Blender test

    A user post on X, dated 26 October 2026, describes an informal Blender test of a model called 'gpt 6.1 terra'.

  • SKY-001 by Skyfall AI

    SKY-001

    A spatial world model for engineering design from Skyfall AI, posted 19 October.

  • JST-2 by Viggle AI

    JST-2

    Viggle's second-generation "physics-native" world foundation model, announced 19 October.

  • PURE by Peking University, HKUST, Pengcheng Laboratory

    PURE

    An ACM Multimedia 2026 paper that post-trains feed-forward 3D reconstruction models on unlabeled video to improve them…

  • Hi3D 1.0 (formerly Sparc3D)

    Hi3D has open-sourced version 1.0 of its image-to-3D model, the renamed Sparc3D, released 23 October 2026.

  • Motion2MixamoRig by MaxLiu22

    Motion2MixamoRig

    An experimental pipeline that takes the motion of one person in a video, or the pose in a still photo, and applies it…

  • InfiniSplat v2 by Zhejiang University, Shenzhen University, Hello TransTech and Causa Robotics

    InfiniSplat v2

    A feed-forward 3D Gaussian Splatting reconstructor that builds scenes from any number of input views, with or without…

  • UniForge3D by Peking University, Math Magic and HKUST

    UniForge3D

    A model, posted 11 October (arXiv 2610.15357), that builds complete 3D assets from sparse, unposed images by…

  • GaussiAnimate

    A SIGGRAPH Asia 2026 method, with its full code released 12 October, that rigs animatable subjects from reconstructed…

  • CoDimRecon by Carnegie Mellon University, Tsinghua SIGS and Genesis AI

    CoDimRecon

    An agentic system that turns multi-view photos into editable scenes ready for physics simulation, modelling soft…

  • DistScene

    Generates complete 3D scenes from images, producing the environment and the objects in it together.

  • GenIA

    A method, posted 8 October (arXiv 2610.12388), that makes Meta's frozen SAM 3D model match real photos and video at…

  • WorldPlay2

    A real-time interactive world model built on Wan2.2 I2V-A14B, steered with frame-aligned actions (camera pitch and yaw,…

  • UniMate by Princeton, UC Berkeley, MIT and NTU

    UniMate

    Since Week 37 2026: preview checkpoints went up on Hugging Face on 27 September, after the training and inference code…

  • Point2Part

    A CMU model that splits a whole 3D shape into parts in one pass, so parts never overlap or leave gaps, with users…

  • PAMI

    Part Anchored Motion for Interaction generates full-body human-object interactions from text.

  • TrackEverything

    A 3D point tracker from CMU and Meta (arXiv, 24 September) that tracks every visible point in long videos, starting…

  • AHa-3D

    A tool harness that has GPT-6 Astra (covered in week 36) turn an ordinary indoor video into an editable, interactive…

  • InSpatio-World 1.5

    Follows InSpatio-World, covered here in week 31 as a 1.3B model that turns a reference video into an explorable 4D…

  • GAE by HKUST, ARC Lab (Tencent IEG), University of Hong Kong & UT Austin

    GAE

    Embeds geometry directly into the generative latent space, so one shared latent decodes to RGB, depth, camera pose and…

  • Mira-Scene by VAST AI Research

    Mira-Scene

    Reconstructs an editable 3D scene from a single image through instance segmentation, depth estimation,…

  • PlayCanvas Engine v2.22.2

    Since Week 19 2026, the engine has shipped three point releases: v2.22.0 on 4 September, v2.22.1 on 8 September and…

  • UMR

    Retargets human motion onto humanoid robots without hand-crafted joint mappings, by learning dense point-cloud…

  • FIRE3D

    Takes an unsegmented image or casual video and does detection, 6-DoF pose estimation and full textured 3D asset…

  • WorldSculpt

    Reconstructs cluttered scenes of hundreds of objects into complete per-object meshes from grounded video, rather than…

  • Marigold V2

    Repurposes a pretrained image-editing diffusion transformer, Qwen-Image-Edit, into single-step dense predictors…

  • Atlas

    Announced 1 September into early access: a multimodal autoregressive diffusion transformer using rectified flow,…