Gemma 3 QAT (Quantization-Aware Training) is Google's latest open-source model, a quantization-optimized version of Gemma 3. Through quantization-aware training technology, Gemma 3 QAT significantly reduces memory usage...
ImagePulse is an open-source project launched by the Moda community, providing dataset support for next-generation image understanding and generative models by atomicating model capabilities and building atomic capability datasets. The project includes multiple...
SimpleAR is a pure autoregressive image generation model jointly developed by the Vision and Learning Lab at Fudan University and the ByteDance Seed team. Employing a concise autoregressive architecture, it achieves high-quality image generation by optimizing the training and inference processes...
Aether is an open-source generative world model from the Shanghai AI Lab, trained entirely on synthetic data. Aether is the first to deeply integrate 3D spatiotemporal modeling with generative modeling, possessing capabilities for 4D dynamic reconstruction, motion-conditional video prediction, and...
Infinite Mobility is an interactive object generation model developed by the Shanghai AI Lab. Based on procedural generation technology, it efficiently generates high-quality interactive object data assets. Infinite Mobility supports 22 common interactive...
ChatTS-14B is a large-scale language model focused on time series understanding and reasoning, open-sourced by ByteDance's research team, with 14 billion parameters. It is finely tuned from Qwen2.5-14B-Instruct and utilizes synthetic data alignment techniques...
MAGI-1 is the world's first autoregressive video generation model open-sourced by Sand AI. It adopts an autoregressive architecture to generate smooth and natural videos by predicting video sequences block by block, and supports infinite expansion and one-shot long video generation.
WriteHERE is an open-source AI-powered long-form writing framework developed by a team led by Jürgen Schmidhuber. WriteHERE is based on heterogeneous recursive planning (HRP) technology, dynamically decomposing the writing task into...
Yuxi-Know is an intelligent question-answering platform built on a large-scale RAG knowledge base and knowledge graph technology. Yuxi-Know supports various knowledge base file formats (such as PDF, TXT, MD, Docx) and supports converting file content...
SurveyGO is an open-source AI-powered academic writing tool developed by the Wallfacer AI team. It supports the rapid generation of well-structured and content-rich review articles based on user-inputted topics and descriptions, helping researchers and students write more efficiently...
Suna, launched by Kortix, is the world's first open-source, general-purpose AI agent project, supporting efficient solutions to various real-world tasks. Based on natural and fluent dialogue interaction, Suna supports intelligent assistance in research and analysis, data processing, and more...
Hunyuan 3D v2.5 is Tencent's next-generation 3D generative model, significantly improving modeling detail. The total number of parameters has increased from 1B to 10B, the number of effective faces has increased more than tenfold, and the effective geometric resolution reaches 1024, achieving ultra-high definition...
AvatarFX is an advanced AI video generation model developed by Character.AI. Based on an uploaded image and selected audio, it instantly brings characters to life, enabling them to speak, sing, and express emotions. AvatarFX supports multiple characters...
Miras is a general-purpose framework from Google for deep learning architecture design, particularly for sequence modeling tasks. Based on the concepts of associative memory and attention bias, Miras redefines models such as Transformer and modern linear RNNs...
SocioVerse is a social simulation world model jointly launched by Fudan University, Shanghai Institute of Innovation and Entrepreneurship, University of Rochester, and Xiaohongshu. It is based on a Large Language Model (LLM) driven intelligent agent and includes 10 million real users...
Pad.ws is an innovative online development environment that combines whiteboard functionality with a full suite of IDE tools. Running in a browser, it requires no additional software installation and can be accessed anytime, anywhere via any device. It integrates an interactive whiteboard with...
GPT-image-1 is a native multimodal image generation model released by OpenAI, available to developers via its API. The model generates high-quality, professional-grade images based on text prompts and images, supporting various styles and customization options...
Eagle 2.5 is a visual language model from NVIDIA, an AI model focused on long-context multimodal learning with only 8 bytes of parameters. Despite its small parameter count, it performs exceptionally well in processing high-resolution images and long video sequences...
Skywork-R1V 2.0 is Kunlun Wanwei's latest open-source multimodal reasoning model, designed specifically for complex reasoning tasks, possessing powerful visual and textual reasoning capabilities. The model is based on a hybrid reinforcement learning and multimodal reward model (Skywork...
Flex.2-preview is an open-source text-to-image diffusion model from Ostris with 8 billion parameters. It supports general control inputs (such as lines, pose, and depth) and has built-in repair functionality. The model can meet a variety of creative needs with a single approach...
DAM-3B (Describe Anything 3B) is a multimodal large language model from NVIDIA, designed to generate detailed descriptions of specific regions in images and videos. The model specifies details using points, bounding boxes, doodles, or masks...
Cooragent is an open-source AI agent collaboration framework developed by the LeapLab team at Tsinghua University. Users can quickly create agents based on simple one-line descriptions, supporting collaboration between agents to complete complex tasks. Cooragent is based on...
Open Avatar Chat is an open-source, modular, real-time digital human dialogue system from Alibaba, supporting full functionality on a single computer. Open Avatar Chat supports low-latency real-time dialogue (average response latency of approximately 2.2 seconds...).
LongPort MCP (Model Context Protocol) is the first brokerage MCP launched by Longbridge Group. Based on LongPort MCP, AI can quickly access securities market data, execute stock trades, manage asset portfolios, etc.