EchoBird is an open-source AI agent desktop management tool developed using Tauri and Rust. The tool integrates over 12 agents, including Claude Code, Codex, and OpenClaw, through the Model Nexus platform...
Grok Imagine Video 1.5 is a graph-based video model from xAI, built on the Aurora autoregressive engine. The model supports one-click generation of short videos with native synchronized audio from static graphs, up to 720p resolution, and more...
AgentCanvas is an open-source Pydantic AI visualization tool from Vstorm. It automatically converts the runtime logs of AI agents integrated with Logfire tracking into an interactive HTML flowchart report, clearly displaying each step...
ACE-Ego is a brain-based multi-morphic embodied manipulation VLA model open-sourced by DaXiao Robotics in collaboration with the Chinese University of Hong Kong's MMLab. The model was pre-trained using over 6.0K hours of human first-person perspective video as its core, and incorporates camera-space motion and morphological conditions...
Image-to-LoRA-V2 (i2L-V2) is an open-source, training-free style transfer tool from the Moda community. It uploads 1-8 style images and predicts the LoRA weights of the raw image model in a single inference.
Qwen-Robot Suite is a foundational model suite launched by Alibaba Tongyi for achieving intelligence in the physical world. It includes Qwen-RobotNav (navigation), Qwen-RobotManip (operation), and Qwen-RobotWorld (world model)...
turbovec is an open-source implementation of Google Research's TurboQuant algorithm. Written in Rust and with accompanying Python bindings, it's a high-performance vector indexing library for RAG scenarios. The tool employs a data-independent quantization strategy...
Seedance 2.0 Mini is a cost-effective and lightweight video generation model launched by ByteDance's Volcano Engine. While retaining core multimodal reference generation capabilities, the model reduces generation costs by approximately 50% and increases speed to Se...
OpenSquilla is an open-source, self-hosted, token-based, high-efficiency microkernel AI agent runtime that emphasizes 'higher intelligence density for the same budget'. It employs a minimalist microkernel architecture (the core has only about 100 lines of code) and utilizes local ML...
Ponytail is an open-source AI Agent code reduction plugin. It injects the minimalist thinking of a seasoned, lazy developer into over 10 mainstream AI coding tools such as Claude Code, Codex, and Cursor, forcing the Agent to simplify its code writing...
Gemini-SQL2 is Google Research's latest AI model, based on Gemini 3.1 Pro, and boasts top-tier Text-to-SQL capabilities. The model achieved 80.04% accuracy in the BIRD benchmark single-model track...
GLM-5.2 is a new generation of open-source large model launched by Zhipu. It has a 1M context and is open-sourced under the MIT license. It focuses on agentic coding and ultra-long context capabilities.
PP-OCRv6 is the sixth-generation open-source OCR model launched by the Baidu PaddleOCR team. It introduces three model sizes for the first time: Tiny (1.5M), Small (7.7M), and Medium (34.5M), covering the entire process from the browser to the server...
openPangu 2.0 is a major upgrade to Huawei's open-source Pangu large model, offering a Pro version with up to 505B parameters and a Flash version with 92B parameters. All models are equipped with a 512K ultra-long context, native depth tuning using Ascend and...
SkillSpector is an open-source AI Agent skills security scanning tool from NVIDIA. It can automatically detect vulnerabilities, malicious patterns, and security risks before skills are installed on platforms such as Claude Code, Codex CLI, and Gemini CLI.
Kimi K2.7 Code is a new generation programming model open-sourced by Moonshot AI. Compared to K2.6, it significantly improves instruction compliance in long-context programming scenarios and performance on long-running tasks, and mitigates overthinking...
html-ppt-skill is an HTML-based AI presentation skill. The tool supports outputting HTML files natively supported by browsers, and includes 36 themes, 31 layouts, 47 animations, and 15 complete templates. (Agenda...)
SwarmFlow is an open-source, multi-agent controllable workflow orchestration framework from openJiuwen. It separates orchestration from intelligence, with collaborative processes executed stably by the system and inference and judgment handled by the Agent.
EvoQuality is a self-evolving visual language model framework jointly launched by ByteDance and City University of Hong Kong, focusing on quality assessment of no-reference images.
Hojo-ASR-V1 is an open-source automatic speech recognition model from the startup team Hojo. It adopts an architecture of Whisper feature extraction + Qwen3-Omni audio encoding + Conformer adaptation + Qwen3-4B language model decoding.
HPC-Ops is an open-source industrial-grade high-performance large model inference operator library from Tencent Hunyuan AI Infra team. It covers core modules such as Attention, MoE, GEMM, sampling and communication fusion, and natively supports BF16/FP8 and mixed precision.
HiDream-O1-Image-1.5 is a commercially viable image generation model launched by Zhixiang Future, based on a native full-modal UiT architecture. It ranks third globally and first in China with an ELO score of 1265 in the Artificial Analysis raw image leaderboard...
MiMo Code is an open-source terminal AI programming assistant developed by Xiaomi's large-scale model team. It's based on OpenCode and licensed under the MIT license. MiMo Code includes the MiMo-V2.5 multimodal model, which is available for a limited time free, and also supports...
SpaceMind is the intelligent space agentic architecture launched by iFlytek, which upgrades space into an L2.5 stage active intelligent agent with the ability to perceive, understand, remember, make decisions and execute.