AI News: 3D, Video, and Language Models for Your Own Hardware
Short news notes about what has changed in the tools: a new version has been released, the license has changed, support for another platform has been added. No press release retelling: if a claim cannot be tied to a repository, model page, or official page, there will be no note.
Articles
134 articles
Spectral Transport Homeostasis method corrects temporal consistency in video withoutA group of researchers published a paper on arXiv about diagnosing and correcting temporal consistency in video generation. The authors introduce the…
Nex-AGI releases Nex-N2.5-mini, a 35B-parameter agentic modelThe model nex-agi/Nex-N2.5-mini with the text-generation task has been published on HuggingFace. It is the smaller model of the new Nex-N2.5 family, f…
MiniCPM5-2B: New 2B Model for Local Deployment Released on HuggingFaceOn September 6, 2026, the openbmb/MiniCPM5-2B model with the text-generation task was published on HuggingFace. It is the second model in the MiniCPM5…
SGLang v0.5.19: Qwen3.8 support, beam search, and DeepEP v2On September 5, 2026, SGLang v0.5.19 was released. The release includes 786 pull requests from 214 contributors. New models include Qwen3.8 (2.4T-A95B…
DSAQuant proposes stage-aware quantization for video diffusionThe authors of an arXiv paper showed that conventional quantization-aware training degrades precisely the details and sharpness of video models, while…
OctWorld presented as a generator of long consistent videos with octree-based 3D memoryOn arXiv, September 3, 2026, a paper on OctWorld was released — a framework for video generation based on diffusion with persistent 3D memory. From a…
Ling-3.0-flash-Fin: Ant Group's 127.5B-Parameter Financial Model Released on HuggingFaceOn September 3, 2026, the inclusionAI/Ling-3.0-flash-Fin model with the text-generation task was published on HuggingFace. It is the first model in th…
LichtFeld-Studio releases LPIPS v0.1 weights based on VGG16LichtFeld-Studio has released LPIPS v0.1 VGG16 weights. This is a persistent model asset in the native .lfw format used by the LPIPS metric evaluator…
Microsoft releases VibeVoice-ASR-Streaming-7B for streaming speech recognitionOn September 2, 2026, the microsoft/VibeVoice-ASR-Streaming-7B model with the automatic-speech-recognition task was published on HuggingFace. The mode…
CC-4DGS reduces dynamic Gaussian Splatting storage to 20–30 MB per sceneA paper titled CC-4DGS has been published on arXiv, proposing a method to reduce the storage footprint of dynamic 4D Gaussian Splatting scenes from te…
K2-Horizon-MoVA-36B-A4B: open-source MoE model with 36B parameters and 4B activeA model IFM/K2-Horizon-MoVA-36B-A4B with the text-generation task has been published on HuggingFace. It is a sparse model of the K2-Horizon family: Mi…
VCAR: Training-Free 3DGS Segmentation via View Completeness and Axis-Aware Boundary RefinementThe authors of the paper VCAR: Training-Free 3DGS Segmentation via View Completeness and Axis-Aware Boundary Refinement propose a method for semantic…
H3-World: Interactive Image-to-Video Model Based on MiniMax-H3The model DANNY621/H3-World with the image-to-video task has been published on HuggingFace. The author describes it as the first interactive world mod…- CapFrame Turns a Text Description of a Frame into a Camera Pose for 3D Gaussian SplattingResearchers presented CapFrame, a method that uses a text instruction to find a 6-DoF camera pose in a 3D Gaussian Splatting scene so that the rendere…
SpatialCrafter turns a single image into an explorable 3D scene via a generative proxyResearchers introduced SpatialCrafter, a two-stage framework for generating explorable 3D scenes from a single image. The work was published on arXiv…
VibeVoice-1.5B-hf: Microsoft's new open text-to-audio model on HuggingFaceThe vibevoice/VibeVoice-1.5B-hf model with the text-to-audio task has been published on HuggingFace. It is designed to generate long conversational au…
4DStreamCtrl combines camera, object, and depth control in streaming modeA paper on arXiv introduces 4DStreamCtrl, a method for interactive video generation that unifies camera motion, object trajectories, and depth into a…
ABot-Recon released on HuggingFace: streaming 3D reconstruction from videoOn August 26, 2026, the acvlab/ABot-Recon model was published on HuggingFace. The authors describe it as a streaming 3D reconstruction model that esti…
LeFlow speeds up planning in world models by an order of magnitude with a generative latent priorResearchers introduced LeFlow, a method that replaces iterative trajectory optimization in latent world models with generative planning. Instead of ru…
Luce introduces 3D asset generation with PBR materials and relighting from a single imageResearchers published on arXiv the paper Luce: Relightable Gaussians for 3D Asset Generation, presenting a method for generating 3D models from a sing…
SceneReGen builds 3D scenes from a single image, generating objects in a shared coordinate systemOn arXiv, August 25, 2026, the paper SceneReGen: Generative Reconstruction of 3D Scenes from a Single Image was published. The authors propose a metho…
GlanceWAM separates imagination and control in a single video DiT for robotsResearchers introduced GlanceWAM, an approach to world-action models that moves video generation off the critical control path. An asynchronous propos…- FixAnything improves 3D scene renders via video generative modelsResearchers introduced FixAnything — a unified model for fixing rendering artifacts across different 3D representations, including Gaussian Splatting…
MiniMax-H3-slim released on HuggingFace for text-to-video generationOn August 24, 2026, the model fkyyy/MiniMax-H3-slim was published on HuggingFace with the text-to-video task. At the time of publication, it has 508 d…
New H3_HD_2K_Detailer model for image-to-video on HuggingFaceOn HuggingFace, on August 24, 2026, the model zuanfilm/H3_HD_2K_Detailer was published with the image-to-video task. At the time of publication, the m…
Loopy turns the DiT temporal axis into a circle for seamless looping videosThe authors of Loopy published a paper on arXiv about generating seamless looping videos. They show that positional embedding layers in DiT control te…
H3 Cinematic Multishot Coverage Released on HuggingFace for Image-to-VideoOn August 24, 2026, the ethanfel/H3_Cinematic_Multishot_Coverage model was published on HuggingFace for the image-to-video task. The code is released…
LTX-Video, CogVideoX and Pyramid Flow fall short of simulators on eight criteriaThe authors of a systematic review on arXiv analyzed 200 papers on generative world models from 2018 to June 2026 and compared them with traditional s…
LTX-Video and CogVideoX Still Don't Solve Multi-Object Generation in a Single VideoA comparative study on arXiv examines three approaches to generating videos with multiple objects without fine-tuning: direct, parallel, and sequentia…
InfinityEdit: Adapter for Infinite Streaming Video Editing UnveiledOn August 21, 2026, the paper InfinityEdit: Infinite Video Editing with a Lightweight Edit-Ignition Adapter was published on arXiv. The authors abando…