AI Briefing#3d-generation
The latest AI and developer news about #3d-generation, with the original source and a short summary.
The latest AI and developer news about #3d-generation, with the original source and a short summary.
The platform includes Viewfinder for scene reconstruction, Chisel for scene shaping, and Splats to Mesh for converting data into game-ready geometry.
A new script and container allow users to fine-tune TRELLIS.2 with LoRA on custom 3D assets, significantly improving geometric accuracy for specific domains like footwear.
The model uses a single network to generate all hierarchical levels (4³, 16³, 64³) from a single image input.
The benchmark tests 145 tasks across 9 categories, finding that GPT-6-astra leads in interactive worlds while local mechanisms outperform stateful workflows.
Gemini 4 Argon leads the benchmark with an Elo of 1544, while Kimi K3.1 Pro offers the highest cost-efficiency at $1.50 total API cost.
Operators of Open3DLab, SFMLab, and SmutBase reported that a Roblox-affiliated crawler bypassed robots.txt and terms of service to scrape 1,027 files totaling 145.5GB, with 62.6GB comprising adult content.
Early testers report that Gemini 4 Carbon offers noticeably better 3D generation than Argon but runs two to four times slower, with frontend and UI capabilities remaining weak.
The benchmark evaluates visual agents' ability to construct 3D scenes in Blender using real-world references.
The Claude Fable 5.1 agent swarm unveiled a project that generates precise 3D worlds of San Francisco and Kyoto using only open data.
World Labs has unveiled Atlas, a next-generation world model that integrates text, image, video, and 3D data.
Apple released the Luce model for 3D asset generation and relighting.
A study has been released on generating programmable 3D objects with built-in animation and logic using LLMs.
An open-source 30B model surpassed GPT-5.5 on the VibeWorlding benchmark, setting a new standard for 3D world generation.
WorldClaw generates editable 3D open worlds from a single prompt.
Tencent has officially released the WorldClaw project based on Hunyuan3D.
image-blaster converts a single image into a 3D environment, objects, and sound effects.
It is an AI agent tool that analyzes a single image and reconstructs it into 3D model code based on Three.js.
Trellis.cpp has launched a Studio feature that supports automatic model downloads and 3D preview.
PixWorld is a unified diffusion model that performs 3D generation and reconstruction simultaneously in pixel space without a VAE.
Image-blaster, an open-source tool that integrates the generation of 3D environments, meshes, and sound effects from a single image, has been released.
TRELLIS.2 was ported to Apple Silicon's MPS, enabling 3D conversion without an NVIDIA GPU.
By combining camera-controlled video generation with 3D reconstruction, it creates persistent, explorable 3D worlds.
HY-World 2.0 has been released, creating editable 3D worlds from text, images, and video.
It generates consistent PBR materials and normals from a single image to support relighting.