Seed Prover 1.5 is a next-generation formal mathematical reasoning model launched by ByteDance's Seed team. The model employs an innovative Agentic Prover architecture, trained through large-scale reinforcement learning (Agentic RL), significantly...
MiniMax M2.1 is a multi-language programming AI model launched by MiniMax. The model systematically improves the capabilities of multiple languages such as Rust, Java, and Golang, covering development from the underlying layer to the application layer.
PromptFill is a structured prompt generation tool designed specifically for AI drawing. Through a visual fill-in-the-blank interaction, it helps users quickly construct complex prompts. It solves the problems of traditional prompts, such as memory overload, grammatical confusion, and poor reusability...
Fun-Audio-Chat is a next-generation end-to-end voice interaction model open-sourced by Alibaba Cloud's Tongyi Bailing team, possessing capabilities in voice understanding, emotion perception, and task execution. The model employs a dual-resolution design, utilizing 5Hz and 25Hz frame rates in tandem...
GLM-4.7 is an open-source model launched by Zhipu AI, focusing on improving coding capabilities, inference abilities, and tool collaboration. The model excels in multi-language programming, complex task planning, and front-end design aesthetics, and supports various programming frameworks...
NitroGen is an open-source, general-purpose game AI model developed by NVIDIA in collaboration with Stanford University, Caltech, and other institutions. It is trained on 40,000 hours of video data covering over 1,000 games and employs large-scale behavior cloning...
Qwen-Image-Layered is an AI image editing model developed by Alibaba. It can automatically split ordinary images into independent RGB transparent layers, achieving layered editing functionality similar to Photoshop. This is achieved through attention mechanisms and positional encoding...
FunctionGemma is a lightweight AI model from Google, focused on optimizing function calls, with 270 million parameters. Based on the Gemma 3 architecture, it is designed to run on mobile devices, browsers, and other edge devices, and can...
T5Gemma 2 is a new open-source edge-side multimodal long context encoder-decoder model from Google. The model optimizes the classic Transformer architecture through regression, introduces innovations such as bound word embeddings and merged attention mechanisms, and features parameter scaling...
GPT-5.2-Codex is an intelligent agent programming model launched by OpenAI, designed specifically for complex software engineering and defensive cybersecurity. The model is an upgraded version of GPT-5.2, improving its capabilities in instruction compliance, long-context understanding, and more...
EcomBench is an AI capability evaluation benchmark for e-commerce scenarios, jointly launched by Tongyi Labs and Skylenage. Built on real-world data, EcomBench covers seven categories including policy consultation, cost estimation, and product selection decisions...
Kairos 3.0 is the first open-source and commercially applicable world model in the industry, based on the ACE embodied development paradigm launched by DaXiao Robotics.
VTP (Visual Tokenizer Pre-training) is an open-source visual generative model pre-training framework from the MiniMax video team. The framework focuses on optimizing the visual tokenizer by integrating contrastive learning, self-supervised learning, and reconstruction learning...
TRELLIS.2 is an open-source 4 billion-parameter 3D generative model from Microsoft, designed for efficiently generating high-fidelity 3D assets from single images. Its core innovation lies in its O-Voxel sparse voxel architecture, capable of handling complex topologies (such as open tables)...
Seed 1.8 is a general-purpose agent model launched by ByteDance, possessing powerful multimodal capabilities and supporting text and image input. The model integrates search, code generation, and GUI interaction capabilities, enabling efficient completion of complex workflows...
SHARP is an open-source AI model from Apple that can convert a single 2D photo into a realistic 3D scene in less than a second. The model uses 3D Gaussian representation technology and a neural network with a single feedforward pass to quickly predict the 3D representation of the scene...
Step-GUI is an AI Agent series model launched by Step-Star, including the cloud-based model Step-GUI and the edge-based model Step-GUI Edge. The cloud-based model excels at handling complex tasks, while the edge-based model is lightweight and can run locally on mobile devices...
SAM Audio is an open-source audio segmentation model from Meta that can separate specific sounds from complex audio mixtures using multimodal cues such as text, visual information, and temporal snippets. Its core technology is the Perception Encoder Audiovi...
Gemini 3 Flash is a high-speed, low-cost, cutting-edge intelligent model from Google. Based on the Gemini 3, it boasts significantly improved performance and powerful reasoning and multimodal understanding capabilities.
Animate Anyone is an open-source framework developed by Alibaba Research Institute of Intelligent Computing that animates characters or people in static images. Based on the diffusion model, this framework incorporates ReferenceNet, Pos...
ReplaceAnything is an open-source AI image content replacement framework developed by Alibaba Research Institute of Intelligent Computing. Utilizing advanced artificial intelligence technology, it can intelligently replace objects in photos or images. Users can...
Outfit Anyone is an open-source project launched by Alibaba Research Institute of Intelligent Computing, offering high-quality virtual try-on for clothing. This technology allows users or models to preview how clothes will look on them without actually trying them on. Outfit Anyone...
PhotoMaker is an open-source model and framework for realistic AI character generation and photo stylization, recently launched by researchers from Tencent PCG's ARC Labs, Nankai University, and the University of Tokyo. It helps users...
FaceChain is an open-source AI framework for generating portraits and digital avatars (similar to a free and open-source version of Miaoduck Camera) launched by Alibaba DAMO Academy. Users only need to provide a minimum of one photo to generate a unique...