SongBloom is a full-length song generation framework developed by Tencent AI Lab. It combines autoregressive sketching and diffusion-based refinement techniques, alternately generating semantic content through an interleaved generation paradigm...
MAI-Image-1 is Microsoft's first self-developed image generation AI model. With a "creator-oriented" design philosophy, the model excels at generating realistic natural lighting effects and complex scene images, such as lightning and landscapes. Compared to some...
Youtu-Embedding is an open-source, enterprise-grade text representation model from Tencent YouTu Lab. Trained on large-scale corpora and using an innovative fine-tuning framework, the model possesses powerful semantic understanding capabilities and is well-suited for handling text...
Kandinsky 5.0 is a text-to-video generation model developed by the Russian AI research lab AI-Forever, boasting powerful generation capabilities and high performance. The core version, Kandinsky 5.0 Video Lite, is...
FS-DFM (Few-Step Discrete Flow-Matching) is a diffusion language model developed by Apple in collaboration with Ohio State University for the rapid generation of long texts.
PySCN is an intelligent code quality analysis tool designed specifically for Python developers. It helps developers quickly identify and clean up code issues such as dead code, duplicate code, and complex code structures. This is achieved through Control Flow Graph (CFG)...
Local-NotebookLM is an open-source local AI tool that converts PDF documents into engaging audio content such as podcasts, interviews, and debates. It supports multiple output formats and content styles, allowing users to choose according to their needs...
Veo 3.1 is Google's latest AI video generation model, a major upgrade from Veo 3, bringing richer audio support, stronger narrative control, and more realistic texture reproduction.
Claude Haiku 4.5 is Anthropic's latest high-performance, low-cost small AI model. Its coding performance is comparable to the previous flagship model, Claude Sonnet 4, and it performs even better in certain tasks...
DiaMoE-TTS is a multi-dialect speech synthesis (TTS) framework jointly developed by Tsinghua University and Giant Interactive. The framework is based on the International Phonetic Alphabet (IPA) unified input system, combined with the dialect-aware Mixture-of-Experts (MoE) framework...
Doubao Voice 2.0 is an upgraded AI voice model launched by ByteDance, comprising three core models: Doubao Voice Recognition Model 2.0 (Doubao-Seed-ASR-2.0), Doubao Voice Synthesis Model 2.0 (Doubao-Seed-TTS 2....
Doubao-Seed 1.6 Lite is a lightweight AI model launched by ByteDance. Compared to the flagship Doubao 1.6, it is lighter, has a faster inference speed, and offers better value for money.
UniPixel is the first unified pixel-level multimodal large model developed by the Hong Kong Polytechnic University and Tencent ARC Lab team, focusing on fine-grained understanding and interaction of images/videos. It can perform object referencing, pixel-level segmentation, and... within a single model.
RTFM (Real-Time Frame Model) is a real-time generative world model developed by Fei-Fei Li's team. The model can run on a single H100 GPU, generating 3D scenes in real time and supporting persistent interaction.
Coral NPU is a full-stack open-source AI platform from Google, designed for low-power edge devices such as smartwatches and AR glasses, addressing three major challenges: performance, fragmentation, and privacy. Coral NPU is based on the RISC-V instruction set...
LongCat-Audio-Codec is an open-source speech codec solution from Meituan's LongCat team, specifically designed for large speech language models (Speech LLM). It utilizes a parallel extraction mechanism of semantic and acoustic dual tokens, taking into account both speech...
PaddleOCR-VL is an open-source multimodal document parsing model from the Baidu PaddlePaddle team. With only 0.9B parameters, it is optimized for low-computing-power devices. It achieved a score of 92.6 in the internationally authoritative benchmark OmnidocBench V1.5, ranking first globally and surpassing GPA...
Dexter is an autonomous financial research agent designed to handle complex financial problems. The tool breaks down complex queries into clear, step-by-step research plans and analyzes them using real-time market data.
DeepSeek-OCR is a visual language model developed by the DeepSeek team, focusing on efficiently processing long text content through optical compression technology. The model consists of a DeepEncoder encoder and a DeepSeek3B-MoE decoder...
KoalaQA is an open-source after-sales service community driven by AI large models, supporting features such as intelligent question answering, semantic search, and automated operations. KoalaQA helps teams achieve 24/7 service, significantly reducing labor costs and improving customer experience...
VitaBench is a benchmark for evaluating large-scale intelligent agents for complex problems, released by Meituan's LongCat team. It uses high-frequency daily life scenarios such as food delivery, restaurant dining, and travel as examples, and has built an interactive system containing 66 tools...
AipexBase is China's first open-source AI-native backend-as-a-Service platform, independently developed by Beijing KuaFuAI. AipexBase aims to enable developers to "achieve AI-native backend service without writing backend code..."
Zen7 Payment Agent, launched by Zen7 Labs, is the world's first decentralized payment agent (DePA). Based on a multi-agent collaborative architecture, it supports multi-chain, multi-currency, high-frequency, small-amount payments, using password-free authorization and gas fees...
HunyuanWorld-Mirror 1.1 is an open-source 3D world generation model released by Tencent. It supports multiple input methods, including multi-view images and videos, and can output point clouds, depth maps, camera parameters, and other 3D geometric data...