Musk's SpaceX AI company announced the launch of a dedicated email feature for the intelligent assistant Grok Bot. This means that the AI assistant is no longer limited to conversational interaction, but now has an independent digital contact point, allowing it to more deeply engage in users' daily tasks and multi-step collaborations. Officially, each AI assistant will be assigned an email address with the suffix mail.grokbot.com. Users can try to customize the prefix of the address, which will be allocated by the system.
OpenAI announced the release of a new Intelligent User Interface on Wednesday, launching alongside GPT-6, marking a major upgrade in ChatGPT's visual and interactive experience. The new interface breaks the limitations of pure text output by integrating frequent visual interaction elements into natural conversation, and introduces highly dynamic components instead of static images. ChatGPT can now directly generate clickable interactive content, enhancing user experience.
Alibaba Qwen Office launched its first desktop robot QwenNote Eva (without screen) at the 2026 Yunqi Conference. It focuses on 'understanding work and understanding you', integrating the core of Qwen Office and the DingTalk ecosystem, claiming that one Eva equals one team. The product is priced at 899 yuan and went on sale that night. It is compatible with Android, iOS, and HarmonyOS. The core emphasis is on 'intelligent collaborative interaction'.
iFLYTEK releases the Spark X2.5 large model, which adopts a Mixture of Experts architecture, with a total of 293 billion parameters and only 30 billion activated during inference, balancing efficiency and performance improvements. It focuses on enhancing code generation and intelligent agent interaction, as well as strengthening general capabilities such as mathematical computation and semantic understanding, and fully supports 256K long context processing.
WhatsApp multi-account warm-up management tool, supporting functions such as automatic conversation and AI intelligent interaction.
Leverage AI to provide multi-channel outreach and intelligent email delivery, and track interactions in real-time.
Intelligent AI customer service agent, providing automated and personalized customer interactions
Puretalk AI? is a multi-channel conversational AI solution that provides intelligent and responsive AI to improve customer interactions across multiple channels.
Openai
$2.8
Input tokens/M
$11.2
Output tokens/M
1k
Context Length
-
Anthropic
$105
$525
200
$7
$35
Google
$2.1
$17.5
$21
$0.7
Alibaba
$2
$20
$4
$16
$1
$10
256
$15.8
$12.7
64
$3.9
$15.2
Bytedance
$0.8
128
$0.15
$1.5
Baidu
32
Xai
$1.4
$10.5
$54
$163
Guilherme34
Qwen3-32B is the latest generation of large language models in the Qwen series, with powerful reasoning, instruction-following, intelligent agent interaction, and multilingual processing capabilities. It supports over 100 languages and dialects and can seamlessly switch between thinking and non-thinking modes, providing users with a natural and smooth dialogue experience.
redponike
Qwen3-VL-4B-Instruct is the latest vision-language model in the Tongyi Qianwen series. It achieves excellent visual perception, text understanding and generation, spatial reasoning, and intelligent agent interaction capabilities at the 4B parameter scale. It supports long contexts and video understanding, and has powerful OCR and multi-language processing capabilities.
unsloth
Qwen3-VL is the most powerful vision-language model in the Tongyi large model series. It has excellent text understanding and generation capabilities, in-depth visual perception and reasoning capabilities, long context support, strong spatial and video dynamic understanding capabilities, and outstanding intelligent agent interaction capabilities. This model uses a Mixture of Experts (MoE) architecture and is a thinking version for enhanced reasoning.
Qwen3 - VL is the most powerful vision - language model in the Tongyi series. It has excellent text understanding and generation capabilities, in - depth visual perception and reasoning abilities, long - context support, strong spatial and video dynamic understanding abilities, and outstanding intelligent agent interaction abilities.
Qwen3-VL-2B-Thinking is the most powerful vision-language model in the Qwen series. It has excellent text understanding and generation capabilities, in-depth visual perception and reasoning abilities, long context support, enhanced spatial and video dynamic understanding abilities, and stronger intelligent agent interaction capabilities. This model uses a 2B parameter scale and supports the instruction version and the thinking version with enhanced reasoning.
quwsarohi
NanoAgent is a compact large intelligent language model with 135 million parameters and an 8k context length. It is specifically optimized for tool calls and instruction following, and can run efficiently on personal devices, supporting lightweight inference and ReAct-style interactions.
Qwen
Qwen3-VL is the most powerful vision-language model in the Qwen series. It has excellent text understanding and generation capabilities, in-depth visual perception and reasoning capabilities, long context support, strong spatial and video dynamic understanding capabilities, and outstanding intelligent agent interaction capabilities. This version uses a mixture-of-experts model architecture and supports enhanced reasoning thinking functions.
geoffmunn
This is the GGUF quantized version of the Qwen/Qwen3-4B language model, a large language model with 4 billion parameters developed by Alibaba's Tongyi series. It is designed for consumer-grade hardware and supports powerful reasoning, intelligent agent workflows, and multilingual interaction.
TEN-framework
The TEN Turn Detection System is an intelligent turn detection model specifically designed for natural dynamic human-machine interaction, accurately identifying natural turn-taking signals to achieve context-aware reasonable interruptions.
Qwen3-8B is the latest large language model in the Qwen series. It has a variety of advanced features, supports multiple languages, and performs excellently in reasoning, instruction following, etc., bringing users a more intelligent and natural interaction experience.
zai-org
Glm-Edge-Chat-4B-GGUF is a text generation model with 4 billion parameters based on the GLM architecture. It is optimized for edge devices and provides an intelligent text interaction experience. The model uses the GGUF format, which is convenient for deployment and operation on various hardware.
HelpingAI
HelpingAI2.5-10B is a language model designed specifically for emotional intelligent dialogue and human-centered interaction. It has a moderate parameter scale but powerful functions, and can provide high-quality services in multiple scenarios.
BrowserTools MCP is a powerful browser monitoring and interaction tool that enables AI to perform intelligent analysis and operations on browsers through a Chrome extension and a local server.
The DevOps AI Toolkit is an AI-based platform engineering and DevOps automation tool that provides functions such as resource deployment recommendations, problem repair, project governance, and a shared prompt library for teams through intelligent Kubernetes operations and conversational workflows, enabling complex cloud-native operations to be achieved through natural language interaction.
Claude Mobile is an MCP server that supports controlling Android, iOS, Desktop, and Aurora OS devices through natural language, providing unified automated operations, intelligent screenshots, permission management, and UI interaction functions.
A Solana wallet intelligent monitoring and trading system based on the MCP protocol, enabling automated operations through natural language interaction with AI.
A Reddit interaction server specifically designed for AI agents, supporting content reading, creation, and sub - forum management, following the MCP protocol and integrating advanced functions such as rule compliance and intelligent replies.
The Unity MCP Server is a C# - based Model Context Protocol (MCP) service used to connect the Unity editor with external large - language models (LLM) or cloud - based AI agents to achieve real - time automation and intelligent interaction.
The FHIR MCP Server is a medical data management server based on the FastMCP framework, providing a complete implementation of the Model Context Protocol (MCP) and supporting the interaction between LLM agents and FHIR-compatible backends. It provides standardized interfaces, supports comprehensive CRUD operations on FHIR resources, and implements intelligent document processing and semantic search through a series of tools.
UIThub MCP Server is a GitHub repository content interaction service designed for Claude, which can intelligently retrieve and analyze code library content through the API.
An intelligent stock data query and analysis service based on the MCP protocol, providing functions such as stock information query, financial analysis, and industry search, and supporting natural - language interaction.
An MCP server for integrating Cursor IDE with Azure DevOps, providing interaction functions for projects, work items, code repositories, and pull requests, and supporting test plan management and intelligent response handling.
An intelligent pet companion application based on the MCP protocol, realizing interaction with virtual pets through voice recognition and natural language processing, and supporting multi - platform operation.
ClaudeHopper is an AI-based construction document assistance tool that enables intelligent interaction with construction drawings and specification documents through RAG and hybrid search technologies.
The DhanHQ trading assistant is an AI-based trading auxiliary tool that realizes stock trading, portfolio management, and account query functions through natural language interaction, supporting various order types and intelligent trading strategies.
An intelligent system that combines AI with Kubernetes management, enabling cluster diagnosis, resource monitoring, and log analysis through natural language interaction to simplify K8s operations and maintenance work.
MCP - Filesystem is a file system operation server optimized for intelligent interaction, providing efficient management of large files and complex file systems, including intelligent context management, precise file operations, and advanced search functions.
MoLing-MineCraft is an AI agent MCP server for the Minecraft game, enabling intelligent building and control through natural language interaction.
AutoMobile is a mobile automation toolset that provides automation testing, UI interaction, test writing, and execution functions for the Android platform. It enables intelligent interaction loops and source code mapping through the MCP server.
An integration toolkit for Wireshark and the MCP protocol, supporting interaction between natural language and network analysis, including a Python server and Lua extension components.
The ElfProxy MCP server is a privacy-first infrastructure solution that combines dynamic IP rotation with AI-optimized network interaction, providing a global proxy IP pool and intelligent web content processing capabilities.
The Xiaohongshu Automatic Search and Comment Tool, developed based on Playwright, supports automatic login, note search, content acquisition, and posting of AI-generated comments. It deeply integrates AI capabilities to achieve intelligent interaction.