JoyAgent-JDGenie is JD.com's first fully-fledged, lightweight, general-purpose multi-agent product open-sourced. As a complete end-to-end intelligent agent system, it can be used directly without secondary development and supports various task processing, such as report generation, ...
Qwen3-Coder is a powerful code generation model developed by the Alibaba Tongyi Qianwen team. It features 480B parameters and 35B activation parameters, supports a native 256K token context, and can be expanded to 1M tokens. The model is used in Agen...
Mureka V7 is an advanced AI music model launched by Kunlun Tech, possessing powerful music creation capabilities. Based on MusiCoT (Music Mind Chain) technology, the model first plans the overall music structure and then fills in the details, generating music that is more...
HunyuanCustom is a multimodal, customizable video generation framework developed by Tencent's Hunyuan team. HunyuanCustom supports various input conditions, including images, audio, video, and text, and can generate videos with specific subjects and scenes...
VoiceCanvas is an open-source multilingual speech synthesis platform. Based on AI technology, it provides high-quality text-to-speech services, supports over 50 languages, and integrates with OpenAI TTS, AWS Polly, and MiniMax, among others...
DeerFlow is an open-source deep research framework developed by ByteDance, enabling users to efficiently complete complex research tasks. DeerFlow combines language models with various tools, such as web search, web scraping, and Python execution, to quickly generate...
WebThinker is a deep research agent proposed by Renmin University of China, Beijing Academy of Artificial Intelligence, and Huawei Poisson Lab, among other institutions. WebThinker empowers large-scale inference models (LRMs) to autonomously perform inference processes...
Multiverse is the world's first AI-generated multiplayer game model, developed by the Israeli team Enigma Labs. It's a multiplayer racing game where players can overtake, drift, and accelerate; each action affects and reshapes the game world in real time. The model...
Open Code Reasoning (OCR) is an open-source code reasoning AI model from NVIDIA, based on the Nemotron architecture, and designed to improve code reasoning and generation capabilities. OCR includes three model versions: 32B, 14B, and 7B, each suitable for...
KuaiMod is a short video quality assessment framework launched by Kuaishou, based on a multimodal large model. It can efficiently identify and filter harmful and low-quality content. The framework draws inspiration from the common law system, using a case-driven approach...
VITA-Audio is an open-source, end-to-end multimodal speech model characterized by low latency and fast inference speed. Through a lightweight multimodal cross-label prediction (MCTP) module, it can generate audio output in the first forward propagation...
Parakeet TDT 0.6B is an open-source Automatic Speech Recognition (ASR) model released by NVIDIA. It employs a FastConformer encoder and TDT decoder architecture, accelerating inference by predicting text tags and their durations, reducing computational complexity...
FunGPT is an open-source project based on the InternLM2.5 series of large models, specifically designed for emotion regulation. It features two core functions: a sweet talk mode and a sharp criticism mode. The sweet talk mode uses warm words and unique...
DMind is a large-scale language model released by the DMind research institute, specifically optimized for the Web3 domain. It is deeply optimized for scenarios such as blockchain, decentralized finance, and smart contracts, fine-tuned using Web3 data, and aligned using RLHF technology.
Dolphin is a lightweight and efficient document parsing model open-sourced by ByteDance. Based on a two-stage approach—parsing the structure first and then the content—the first stage generates a sequence of document layout elements, and the second stage uses these elements as anchor points...
Playwright MCP is a lightweight browser automation tool from Microsoft, based on the Model Context Protocol (MCP). The tool interacts with web pages based on Playwright's accessibility tree, without relying on...
RelightVid is a temporal consistency diffusion model for video relighting, developed by Shanghai AI Lab, Fudan University, Shanghai Jiao Tong University, Zhejiang University, Stanford University, and the Chinese University of Hong Kong. It supports rendering based on text cues, background visuals, etc.
HRAvatar is a monocular video reconstruction technology developed by Tsinghua University in collaboration with the IDEA team. It supports the generation of high-quality, relit 3D avatars from ordinary monocular videos. HRAvatar uses learnable deformation bases and linear skinning techniques...
Pocket Flow is a minimalist LLM (Large Language Model) framework, implemented in just 100 lines of code. It is lightweight, dependency-free, and vendor-unlocked. Pocket Flow supports multiple agents, workflows, retrieval enhancements, and more...
Aurora is a 1.3 billion parameter atmospheric model developed by Microsoft Research. It extracts valuable information from massive amounts of atmospheric data to predict global weather patterns, air pollution, and atmospheric processes such as ocean waves. The model uses pre-trained...
DeepWiki MCP is a remote server developed by Cognition Labs, based on the open standard Model Context Protocol (MCP). DeepWiki MCP provides AI applications with access to and search of GitHub codebase documentation...
Gemini Diffusion is an experimental text diffusion model introduced by Google. Unlike traditional autoregressive models that generate text word by word, Gemini Diffusion generates output by progressively refining noise, enabling rapid iterative error correction.
3DTown is a framework developed by Columbia University in collaboration with Cybever AI and other institutions to generate 3D town scenes from a single top-down view. The framework is based on region-based generation and spatially aware 3D inpainting techniques, decomposing the input image into overlapping regions...
Sparc3D is a high-resolution 3D model generation framework developed by Nanyang Technological University in collaboration with Sensory Universe and Imperial College London. It addresses the issues of detail loss and inefficiency in traditional 3D generation methods. The framework combines sparse...