DeepTutor is an open-source AI learning assistant from the Data Science Lab at the University of Hong Kong. It transforms complex knowledge systems into an efficient learning experience through a multi-agent architecture and knowledge graph technology.
Nemotron Speech ASR is an open-source model from NVIDIA focused on low-latency, real-time streaming speech recognition. Through a cache-aware architecture, it caches processed speech features and only performs calculations on new audio frames, achieving single-sentence transcription...
FantasyWorld is an innovative 3D world modeling framework developed in collaboration between AMAP (Amap) and Beijing University of Posts and Telecommunications. It focuses on generating high-quality 3D scenes through unified video and 3D prediction. The framework uses frozen video as a base...
Qwen3-VL-Embedding is a multimodal information retrieval model launched by Alibaba Tongyi, designed specifically for handling various modal inputs such as text, images, visual documents, and videos. Based on the powerful Qwen3-VL architecture, the model can...
Qwen3-VL-Reranker is a cross-modal understanding model built by Alibaba Tongyi based on Qwen3-VL, specifically designed for multimodal information retrieval. The model accepts query-document pairs with arbitrary modal combinations (e.g., matching text-image queries with text-document pairs...).
VoiceSculptor is a voice design model developed by Northwestern Polytechnical University, YTO Intelligence, and other institutions. It can achieve fine-grained control over speech synthesis through natural language commands.
ChatDev 2.0 is a no-code multi-agent tool open-sourced by a joint team from Tsinghua University, Shanghai Jiao Tong University, OpenBMB, and Wallfacer Intelligence. The tool uses a visual interface, allowing users to quickly build multi-agent systems through drag-and-drop configuration...
EvoCUA (Evolving Computer Use Agent) is an open-source multimodal model from Meituan, focusing on automating computer usage tasks. EvoCUA uses natural language commands and screenshots to automate tasks in Chrome, Excel, Power...
MMSI-Video-Bench is a benchmark tool for evaluating the capabilities of Multimodal Large Language Models (MLLMs) in video spatial intelligence. Developed jointly by the Shanghai Artificial Intelligence Laboratory and several universities, it comprehensively evaluates models in real-world scenarios...
Youtu-LLM is a lightweight language model open-sourced by Tencent's Youtu team, with 1.96 billion parameters. Designed specifically for agent tasks, it possesses powerful "native agent capabilities," outperforming models of similar or even larger scales in multiple tasks...
OS-Copilot is an open-source operating system intelligent assistant project aimed at building a general-purpose computer intelligent agent capable of interacting with various elements within the operating system, such as networks, code terminals, files, multimedia, and third-party applications. Its core is FRI...
The 10Kh RealOmni-Open dataset, open-sourced by the Gen Robot.AI team, is one of the largest open-source datasets in the industry. It contains over 10,000 hours of data, more than 1 million task clips, and a total storage capacity of...
UI-TARS Desktop is an open-source desktop automation tool developed by ByteDance, based on a visual language model. It supports controlling computer operations via natural language commands, such as opening files, browsing web pages, and operating software, and can accurately recognize screen input...
Open Interpreter is an open-source AI endpoint assistant project that provides native code execution capabilities for large language models (LLMs). Through a natural language interface, users can interact with... in a manner similar to ChatGPT within the endpoint.
TeleChat3 is a fine-grained MoE semantic model with hundreds of billions of parameters, independently developed by the Telecom Artificial Intelligence Research Institute (TeleAI) of China Telecom. It is trained using domestically produced computing power and supports tasks such as multi-turn dialogue, text creation, and code generation. The model package...
openPangu-VL-7B is an open-source multimodal model launched by Huawei, optimized specifically for Ascend hardware. Combining language and vision capabilities, the model possesses powerful visual localization and OCR functions, enabling efficient processing of image, document, and video tasks.
MiroThinker v1.5 is an open-source search agent model from the MiroMind team. Through interactive extension techniques, the model deeply couples inference with the external environment, breaking away from the limitations of traditional large models that rely on massive parameters.
TuriX-CUA is an open-source AI agent based on Python that automates desktop operations using AI technology. TuriX-CUA "sees" the screen by taking screenshots, uses a multimodal large model to "think" about the next action, and completes user-instructed tasks...
DLCM (Dynamic Large Concept Models) is a new large language model architecture launched by the ByteDance Seed team. The framework elevates the model's reasoning unit from the traditional token (word) level to the concept (concept) level, through...
Paper2Any is an open-source, multimodal AI-assisted platform developed by the DCAI research group at Peking University. Based on the automated data governance agent framework DataFlow-Agent, it can automatically convert research papers into editable PPT and SVG files. ...
OpenCode is an open-source AI programming assistant that improves development efficiency and code quality. The tool exists as a terminal interface, desktop application, or IDE plugin, and integrates with various Language Model (LLM) providers to provide development...
StoryMem is an open-source video generation framework jointly released by ByteDance and Nanyang Technological University, addressing the long-standing cross-shot consistency problem in AI video generation. It utilizes an explicit visual memory mechanism to diffuse single-shot video...
Voquill is an open-source voice input tool that replaces traditional keyboard typing with voice input, improving writing and note-taking efficiency. It supports macOS, Windows, and Linux systems and can be used in any text box and application, enabling...
XVERSE-Ent is an open-source, bilingual (Chinese and English) large-scale model designed specifically for the pan-entertainment industry by Yuanxiang Technology. It includes the Chinese model XVERSE-Ent-A4.2B and the English model XVERSE-Ent-A5.7B. The models emphasize character consistency, long storylines, etc.