MiniCPM-Robot is a lightweight embodied intelligence (VLA) model series from Wallfacer Intelligence, including the MiniCPM-RobotManip general operation model with 1.5B parameters and the MiniCPM-RobotTrack tracking and navigation model with 0.9B parameters...
Qwen-Audio-3.0-TTS is a large-scale speech synthesis model launched by Alibaba's Tongyi Qianwen, which includes a Flash version for real-time interaction and a Plus version for high-quality generation.
LoHoSearch is a next-generation search intelligence benchmark launched by Meituan's LongCat team. It automatically generates questions based on a knowledge graph covering 7.62 million Wikipedia entities and 265 million relationship edges, replacing traditional manual question generation...
Vivago R1 is the world's first unlimited-duration multimodal creative intelligent agent launched by Zhixiang Future. Based on its self-developed HD-AgentOS, it achieves multi-agent collaborative scheduling, supports the continuous generation and editing of videos of any length, and possesses long-task capabilities...
Matrix-Game 3.5 is an open-source real-time streaming interactive world model developed by the Kunlun Wanwei Riemann Power team. It achieves three-dimensional spatial long-term memory and geometric consistency modeling based on Patch Memory and Warped ProPE.
Tempolor v4.7 is a flagship AI music generation model launched by Quwan Technology. It adopts a fourth-generation layered progressive architecture and supports 48kHz dual-channel high-quality output.
Qwen3.8-Max-Preview is a preview version of the latest generation of the base model launched by Alibaba's Tongyi Qianwen, with 2.4T of parameters. It has been launched on the Qoder, Token Plan and QoderWork platforms.
MuScriptor is an open-source multi-instrument music transcription model jointly developed by Kyutai and Mirelo, which can automatically transcribe real-world music audio from various genres into MIDI.
StaffDeck is an open-source digital employee end-to-end workflow building and management platform developed by Facewall Intelligence in collaboration with the Northeastern University-Facewall Intelligence Data Intelligence Joint Laboratory, Tsinghua University THUNLP Laboratory, OpenBMB, and AI9Stars.
Wan-Streamer v0.2 is an end-to-end, full-modal understanding and generation model for real-time duplex interaction, developed by Alibaba's Tongyi Lab. The model unifies 'listening, seeing, speaking, and acting' into a single Transformer, natively supporting text, ...
Kimi K3 is an open-source model with 2.8 trillion parameters launched by Dark Side of the Moon. It is built on the KDA hybrid linear attention mechanism and attention residual technology, natively supports visual understanding, and has a context window of 1 million tokens.
Inkling is an open-weighted multimodal foundational model from Thinking Machines Lab. The model natively supports text, image, and audio input, employs a hybrid expert architecture, and achieves cross-modal inference with 41 billion activation parameters...
Colibrì is an open-source, lightweight local inference engine capable of running the flagship GLM-5.2 MoE model with 744 parameters on a consumer-grade computer with 25GB of RAM. The model uses a hierarchical storage architecture to keep resident parameters in memory...
Xiaomi-Robotics-U0 is a unified embodied synthesis model with 38 billion parameters launched by Xiaomi. It is continuously trained based on a world-based model and jointly optimizes text-to-image, image editing, embodied scene generation, embodied transfer, and embodied video generation...
X2.0 is the world's first real-time interactive video generation model launched by Xmax AI, supporting millisecond-level streaming generation. Users can use their cameras to change character outfits, summon virtual objects, and control the screen in real time, achieving playable video...
Qwen-Audio-3.0-Realtime is a real-time voice interaction dialogue model developed by the Alibaba Tongyi team, offering both Plus and Flash versions. The model maintains high inference depth while achieving millisecond-level response times and supports instructionless agent...
ABot-World Studio is a universal world model workshop launched by AutoNavi, unifying interactive video generation and 3DGS scene generation into a single product. Users can input text or images to generate an AI world that can interact in real time, supporting...
OvisOCR2 is an end-to-end document parsing model developed by Alibaba's ATH-MaaS team. It is trained on Qwen3.5-0.8B and is completely open source. The model topped the OmniDocBench v1.6 benchmark with a score of 96.58, demonstrating its ability to...
MOSS-VL-Realtime is an open-source 11-parameter streaming visual language model from OpenMOSS, designed specifically for real-time video understanding. The model supports on-the-fly responses, real-time corrections, and active silencing, upgrading video understanding from simply watching recordings to...
StepAmoo is a new generation of personal intelligent agent launched by Step AOS, based on the Step AOS intelligent agent native operating system, possessing an operating system-level identity. StepAmoo has five core capabilities: perception, memory, planning, connection, and execution...
AnySearch is a real-time structured search engine designed specifically for AI agents, accessible via API, MCP, and Skill. The product covers general web pages and data sources from over 20 vertical sectors, including code, law, ...
PixVerse Game is the first real-time video game engine launched by iSpeed Technology. Built on the PixVerse R1 real-time world model, it constructs a complete game experience on a single real-time interactive video stream, allowing creators to define...
HyOCR-1.5 is a lightweight, end-to-end OCR expert model open-sourced by Tencent Hunyuan. With only 1B of parameters, it can perform document parsing, text recognition, information extraction, image translation, chart parsing, ancient script recognition, video subtitle extraction, and more...
SenseNova-Vision is a unified vision model open-sourced by SenseTime, which natively integrates classic vision tasks such as object detection, image segmentation, depth prediction, and 3D reconstruction into the large model base.