AI tools for Open Source

AirMusic is an innovative platform leveraging advanced algorithms to allow users to generate completely original music in a matter of seconds. By simply providing input regardin

DeepSeek Harness is an open-source runtime for coding and general-purpose agents. It is available in public preview as a desktop application and as a developer-facing web runtim

Ripple-65B is Ricursive Intelligence's open-weight language model for general-purpose reasoning and agentic coding. It is a dense 65-billion-parameter transformer derived from t

ERNIE 5.5 Flash Base is Baidu's released foundation checkpoint for further model training. The sparse language model has 280 billion total parameters and activates approximately

OpenRoboKit is a suite of open-source libraries for robotics development. Its components cover motion optimization, transfer of motion across embodiments, manipulation and agent

Cactus Mirage is a compact vision model for local image understanding and structured extraction. It can describe a scene, locate objects, read text and produce masks or structur

ReLift reconstructs dense 3D point trajectories from a monocular video. It combines the output of a 2D tracker with a video geometry estimator to recover how points move in spac

ToolSpace is a research toolkit for turning tool-use traces into executable training and evaluation environments. It focuses on long-horizon agent tasks that require several sta

RelayVSR performs 4x super-resolution on real-world video. It uses a large generative model selectively to recover detail and a smaller network for the rest of the sequence.

CoLa3D reconstructs an object-composed 3D scene from a single image. It outputs individual meshes that share a coordinate frame so the resulting objects can be inspected and man

P-JiT 21K is Linum's released research checkpoint for text-to-image generation. The two-billion-parameter model was trained on 660 million samples to investigate what a smaller

FlowHMR reconstructs 3D human motion from monocular video. It targets motion that can be tracked in a physics simulation and provides tools to preview the recovered sequence and

Hyra Results is Tencent Hunyuan's repository of released scientific models and reproducible research artifacts. Its model directories provide usable packages and checkpoints for

STRATA is NVIDIA's global atmospheric emulator for research on storm-resolving simulation. It learns to predict the evolving three-dimensional atmospheric state represented by t

P-H3-Image is Pruna AI's open-weight image generator and editor derived from MiniMax-H3. It supports text-to-image work and edits conditioned on several reference images, with a

NitroFusion2 is a text-to-image checkpoint designed to render a result in one or two diffusion steps. It builds on Z-Image-Turbo and targets sharp visual detail, including text

HairDeck turns a character image into stylized 3D hair cards for animation and avatar workflows. It produces a set of textured strips that follow the overall silhouette and the

This Qwen3.5-0.8B checkpoint is the released result of a weight-space programming experiment. The authors tasked an external language model with modifying the smaller model's te

Hydrogen-1.0 is Deep Principle's language model for scientific problem solving. It is built on GLM-5.3 and targets tasks such as explaining derivations, organizing research ques

Xiaomi-Robotics-1 is a vision-language-action foundation model for mobile manipulation. Xiaomi trained the model on more than 100,000 hours of real-world manipulation trajectori

SenseNova-Loop-9B-Preview is an open multimodal language model for long-context and agentic workflows. It accepts text and images and generates text, supporting document interpr