Amodal3R is a conditional 3D generative model that infers and reconstructs the complete 3D shape and appearance from partially visible 2D object images. The model is built upon the "base" 3D generative model TRELLIS, incorporating mask weighting...
AutoAgent is a no-code, automated LLM (Learning Model) intelligent agent framework developed by the University of Hong Kong. Based on natural language interaction, it allows users to create intelligent assistants without programming, suitable for scenarios such as intelligent search, data analysis, and report generation...
Text to Bark is the world's first AI-powered text-to-speech model for dogs, developed by ElevenLabs. Users input text and select the dog breed, and the model generates highly realistic dog barks that are indistinguishable from real barks in 95% of dogs. ...
Cua is an open-source AI agent project launched by the trycua team, providing macOS users with high-performance virtualization and AI agent capabilities. Cua is based on Apple's Virtualization Framework and supports Apple Silice...
SWEET-RL is a multi-turn reinforcement learning framework from Meta, specifically designed for training large language model (LLM) agents for collaborative reasoning tasks. SWEET-RL optimizes the "critic"... based on additional information during training (such as a reference solution).
OThink-MR1 is a multimodal language model optimization framework jointly developed by OPPO Research Institute and Hong Kong University of Science and Technology (Guangzhou). Based on dynamically adjusting the Kullback-Leibler (KL) divergence policy (GRPO-D) and reward model, it improves...
DeepSite is an online development tool based on the DeepSeek-V3 model. Users do not need to configure the environment or install software. They can directly input their requirements on the webpage and quickly generate code for games, applications, or webpages, supporting real-time preview.
EasyControl Ghibli is an AI model developed based on the EasyControl framework. It's available on the Hugging Face platform and focuses on converting ordinary images into Ghibli-style images. It was created using only 100 photos of Asian faces...
Dolphin is a large-scale speech model for Eastern languages, jointly developed by the Speech and Audio Technology Laboratory of the Department of Electronic Engineering at Tsinghua University and Haitian Ruisheng. It supports speech recognition for 40 Eastern languages, including 22 dialects of Chinese (including...).
WorldScore is a unified evaluation benchmark for world generation models proposed by Stanford University. It decomposes world generation into a series of next-scene generation tasks, achieving different levels of performance through explicit camera trajectory-based layout specifications...
PaperBench is an open-source AI agent evaluation benchmark from OpenAI, supporting the assessment of an agent's ability to reproduce top-tier academic papers. PaperBench requires agents to comprehensively demonstrate their capabilities, from understanding the paper's content to writing code and executing experiments...
DreamActor-M1 is an advanced AI image animation framework launched by ByteDance, capable of transforming static portrait photos into vivid animated videos. It employs a hybrid guidance mechanism, combining implicit facial representation, 3D head spheres, and 3D body skeletons, etc...
MoCha AI is an end-to-end dialogue-based character video generation model jointly developed by Meta and the University of Waterloo. It can generate complete character animations with synchronized speech and natural movements based on text or speech input. MoCha employs a speech-view...
DeepSeek-GRM is a generalist reward modeling system jointly proposed by DeepSeek and researchers from Tsinghua University. It utilizes pointwise generative reward modeling...
OlympicArena is a multidisciplinary cognitive reasoning benchmark framework jointly developed by Shanghai Jiao Tong University, Shanghai AI Lab, Soochow University, and the Generative Artificial Intelligence Laboratory (GAIR Lab) of Shanghai Jiao Tong University. OlympicArena includes...
AbletonMCP is an open-source project that connects Ableton Live and Claude AI based on the Model Context Protocol (MCP) to enable AI-assisted music production. AbletonMCP is based on bidirectional communication and supports users to...
BabelDOC is an open-source intelligent PDF translation tool designed specifically for translating scientific papers. It generates translated text alongside the original, creating a bilingual comparison without switching windows, making it easy to read. It fully preserves mathematical formulas, tables, and...
Llama 4 is a multimodal AI model series from Meta, an open-source developer. It is the first to adopt a hybrid expert (MoE) architecture, resulting in higher computational efficiency during training and inference. Llama 4 currently has two versions: Scout and Maverick. Scout...
OmniTalker is a real-time text-driven speaking avatar generation technology released by Alibaba. It can simultaneously process multiple modalities such as text, images, audio, and video input, generating natural speech responses in a streaming manner. Its core architecture is based on T...
Seedream 3.0 is an AI image generation model launched by ByteDance. The model supports native 2K resolution image output and quickly generates high-quality images in just 3 seconds. The model excels in small text generation and layout, aesthetics, and...
Quasar Alpha is a pre-release AI model supporting a massive context window of 1 million tokens, capable of handling extremely long texts and complex documents. It boasts excellent code generation capabilities, with fast generation speed, low latency, and strong instruction adherence...
OmniCam is an advanced multimodal video generation framework that achieves high-quality video generation through camera control. It supports multiple input modal combinations, allowing users to provide text descriptions, video trajectories, or images as references...
AnimeGamer is an infinite anime life simulation system jointly developed by Tencent PCG and City University of Hong Kong. Based on a multimodal large language model (MLLM), it allows players to immerse themselves in anime life using open-ended language commands, embodying the characters of anime...
Awesome MCP Servers is an open-source project that aggregates various servers based on the Model Context Protocol (MCP). Awesome MCP Servers compiles over 3000 MCP servers, covering browser auto-navigation...