GLM-4-Voice is an end-to-end emotional speech model developed by Zhipu AI. It can directly understand and generate Chinese and English speech, support real-time voice dialogue, and flexibly adjust the emotion, tone, speed, and dialect features of the speech according to user commands.
Unbounded is an infinite life simulation game co-developed by Google and the University of North Carolina at Chapel Hill. Breaking away from the limitations of traditional video games, the game utilizes generative models, such as Large Language Models (LLM) and visual generative models, to create...
AutoGLM, launched by Zhipu, is the world's first mobile agent, providing users with efficient operational proxy services through AI technology. Based on the domestically developed GLM-4.5 and GLM-4.5V models, AutoGLM possesses inference, coding, and multimodal capabilities. Aut...
EveryoneNobel is an open-source AI tool that generates personalized Nobel Prize-style images for everyone. Based on the ComfyUI framework, EveryoneNobel combines HTML templates and image generation technology; users simply upload a portrait photo...
TextHarmony is a multimodal generative model jointly developed by East China Normal University and ByteDance, adept at understanding and generating visual text. The model is based on Slide-LoRA technology, dynamically aggregating modality-specific and modality-independent LoRA experts...
BlinkShot is a real-time AI image generator that rapidly produces high-quality images. Users simply input prompts, and BlinkShot generates an image within milliseconds. The tool is based on Together AI's Flux Schnell technology and supports...
NotebookLlama is an open-source project from Meta that converts PDF documents into podcast content. The project is based on a series of automated steps, using the LLaMa model for PDF preprocessing, generating podcast scripts, adding dramatic elements, and more...
WonderWorld is an innovative 3D scene generation framework jointly developed by Stanford University and MIT. It can quickly generate diverse and coherent 3D virtual worlds from a single image. It is based on the core Fast Layered Gaussian...
Meissonic is a text-to-image synthesis model developed in collaboration with Alibaba Group, Skywork AI, and several universities. Based on masked image modeling techniques, it combines multimodal and unimodal Transformer layers, high-level positional encoding strategies...
Zamba2-7B is a small language model from Zyphra, featuring an innovative architecture that achieves fast inference speed and low memory usage while maintaining output quality. The model performs exceptionally well in tasks such as image description and is suitable for...
PersonaTalk is a two-stage attention-based framework developed by ByteDance for achieving high-fidelity and personalized visual voice-overs. PersonaTalk can synthesize videos with precise lip-sync to the target audio while...
OpenSPG is a knowledge graph engine based on the SPG framework, launched by Ant Group in collaboration with the OpenKG community. OpenSPG integrates the structure of LPG and the semantics of RDF, overcoming the challenges of implementing complex semantics in RDF/OWL, and inheriting the structure of LPG...
LongVU is a long-video understanding model developed by the Meta AI team, based on a spatiotemporal adaptive compression mechanism. It addresses the challenge of processing long videos limited by the context size of large language models (LLMs). LongVU is based on cross-modal queries and...
SynthID Text is a text watermarking technology developed by Google DeepMind, used to identify and verify text generated by large language models (LLMs). It relies on subtle adjustments to the token probability score embedding during the generation process, making it virtually imperceptible...
VILA-U is a unified foundational model integrating video, image, and language understanding and generation. It handles both understanding and generation tasks based on a single autoregressive next-label prediction framework, simplifying the model structure and excelling in visual language understanding and generation...
Video-XL is an ultra-long visual understanding model designed specifically for understanding hour-long videos, jointly developed by researchers from the Beijing Academy of Artificial Intelligence, Shanghai Jiao Tong University, Renmin University of China, the Chinese Academy of Sciences, Beijing University of Posts and Telecommunications, and Peking University. Based on...
Embed3, developed by Cohere, is an industry-leading multimodal AI search model that generates embedding vectors from text and images, helping businesses quickly and accurately search complex reports, product catalogs, and design documents—multimodal assets in general.
DriveDreamer4D is a framework for improving the quality of 4D reconstruction in autonomous driving scenarios. It enhances the representation of 4D driving scenarios based on a prior world model. The framework can synthesize new trajectory videos based on real-world driving data, using explicit structures...
Animate-X is a general-purpose animation framework based on LDM (Laser Dynamics Model), capable of converting still images into dynamic videos, and excels at handling anthropomorphic characters. By introducing pose indicators, it enhances its ability to capture motion patterns, including implicit and explicit movements...
MarDini is a novel video diffusion model that combines the advantages of Mask Autoregressive (MAR) and Diffusion (DM) models for large-scale video generation. The model can handle any number and location of masked frames and supports video interpolation, image...
SDXL-EcomID is an open-source project launched by Alimama that generates customized personalized images from a single reference image. It combines the advantages of PuLID and InstantID, optimizing background consistency, facial landmark control, and facial realism to improve...
DreamClear is a high-performance image restoration technology jointly developed by the Institute of Automation, Chinese Academy of Sciences and ByteDance. It focuses on privacy-preserving dataset management and can restore low-quality (LQ) images to high-quality (HQ) images.
GitHub Spark is an AI programming tool launched by GitHub, supporting four programming models including GPT-4o and Claude Sonnet 3.5. Users don't need any code or deployment environment; they can describe their needs using natural language and run the program on desktop or mobile devices...
SimpleQA is a benchmark developed by OpenAI to evaluate the ability of large language models to answer short, fact-finding questions. SimpleQA contains 4326 questions, each designed to have only one correct answer, making it easy to score...