Back to News
SourceMarsBit

xAI面向Grok推出Build模型

火星财经消息,xAI面向Grok推出Build模型。Grok Build模型处于早期测试阶段,针对的是SuperGrok Heavy超级订阅者。(财联社)
Disclaimer: The views above are the author's only and do not represent 711BTC. Nothing here constitutes investment advice.

Related

05-26 08:58

xAI's Grok Build beta is now available to SuperGrok and X Premium+ users.

xAI announced on May 25th (local time) that Grok Build is now open for beta testing to all SuperGrok and X Premium+ users. This feature supports Plan Mode, generating images and videos via the Imagine tool, and building automated workflows or orchestrators through the command-line interface (CLI). (Interface)

05-18 11:26

xAI launches price war, aggressively promoting Grok Build, dropping from $300 to $99.

According to Beating, xAI has launched an aggressive limited-time subsidy strategy for its premium subscription service, SuperGrok Heavy. New and existing users can enjoy a massive discount of 67% from $300/month to $99/month over the next six months. The Heavy subscription is currently the only barrier to entry for xAI's command-line programming tool, Grok Build. This tool, which launched its public beta last week, focuses on concurrent subagents and native MCP extensions.

06-02 14:59

Grok Build integrates with Composer 2.5: xAI terminal agents introduce the Cursor main model.

According to Beating, xAI, Elon Musk's AI startup, announced that its command-line agent development tool Grok Build has fully integrated Composer 2.5, a top-tier programming model developed by Cursor. Developers subscribing to X Premium+ or SuperGrok can directly switch to Composer 2.5 in the terminal menu. Cursor released Composer 2.5, a high-performance programming model, on May 18th, focusing on high-precision multi-file editing and long-term instruction compliance. The model's underlying architecture uses the commercially licensed open-source Kimi K2.5 from Dark Side of the Moon, and was trained using large-scale reinforcement learning and synthetic task distillation by the Cursor team. The "local reinforcement learning with text feedback" technique introduced during training provides targeted hints at precise points where tool calls fail. The model can automatically correct local errors, improving the success rate of cross-file code refactoring and long-term execution. Official demonstrations show that when faced with natural language commands such as "improve the mobile responsiveness of the game loading menu," the agent can autonomously retrieve relevant code. After performing regular expression searches on files such as useMobile.ts and Game.tsx, the agent directly locates and modifies the specific implementation of localStorage in page.tsx. The interface supports shortcut commands such as `/fork` (clone the current session to a parallel sub-agent), `/compact` (compress history to reduce token consumption), and `/copy` (copy the latest answer), and provides an "always-approve" automatic authorization mode. The terminal ecosystem integration stems from the computing power sharing cooperation between Cursor and SpaceXAI. Currently, Cursor has access to computing resources, including the Colossus 2 super cluster, for model training. The introduction of Composer 2.5 is the first implementation of the cooperation between the two parties in the xAI terminal tool.

06-12 11:03

xAI launches Grok Build built-in plugin marketplace: one-click expansion of AI terminal skills, and forced SHA security lock.

According to Beating, xAI has announced the launch of the Grok Build Plugin Marketplace, a built-in plugin marketplace for its endpoint AI programming tool, Grok Build. This allows developers to directly browse, install, and update third-party tools within the endpoint. Each plugin integrates and packages modules such as the AI's underlying skills, control protocols (MCPs), and code completion tools (LSPs), eliminating the hassle of fragmented configuration. Since the marketplace is directly integrated into the command-line interface, users can update and iterate plugins without leaving the endpoint. Among the initial plugins launched, the MongoDB plugin allows AI to directly query and optimize databases, the Vercel plugin supports control over cloud deployments and domains within the endpoint, the Sentry plugin helps AI analyze error stacks to automatically fix bugs, and the Chrome DevTools plugin allows AI to control the browser in real time and record performance tracks. In addition, the Cloudflare developer skill pack and the community workflow tool Superpowers have also been launched simultaneously. Users can access the directory by typing `/marketplace` in Grok Build or by running `grok plugin install` via the command line.

05-30 14:15

xAI accelerates the iteration of Grok Build, adding X search, file parsing, and multi-agent capabilities, moving towards a mature AI programming platform.

According to BlockBeats, on May 30th, xAI has been consistently updating its AI programming tool, Grok Build, at a high frequency. Version 0.2.11 has introduced several new features and performance optimizations, rapidly evolving from an early CLI tool into a complete Agentic Coding environment. Functionally, Grok Build integrates X platform search and faster web search capabilities, and adds commands such as `/export`, `/login`, `/usage`, and `/config-agents`. It also supports an interactive file reader and PowerPoint text extraction. Furthermore, the platform introduces an Always-approve mode to simplify the permission authorization process. In terms of compatibility, Grok Build adds support for Windows ARM64 and macOS x86_64, optimizes the experience in various development environments such as Warp, JetBrains, and traditional Windows terminals, and fixes copy-paste issues on Linux Wayland, WSL, and Windows platforms. Regarding Agent capability upgrades, subagents can now share the terminal backend, task scheduler, and monitoring system across sessions. A proactive reminder mechanism and a "lazy detector" have also been added to improve model task execution efficiency. Furthermore, the system has further optimized context compression and chat history memory management, and supports automatic background execution of long-running Bash commands. In terms of user experience, the new version increases the terminal video playback frame rate to 30FPS, adds multi-image pasting, drag-and-drop upload, and temporary screenshot functionality for macOS, and optimizes details such as model switching loading prompts, Markdown link rendering, and editor line number display. In terms of stability, xAI increases the default retry budget to approximately 5 minutes, improves Unix system resource limits (ulimit), and enhances the handling of timeouts and abnormal terminations in background tools. It also fixes several compatibility issues related to UTF-8 output, large-screen displays, and Windows high-contrast mode. --------------------------------- Click the original link below to join the Beating · Lark AI news channel and monitor global AI hot topics and news 24/7.

07-01 23:40

xAI releases Voice Agent Builder, a voice intelligence platform: enabling the creation of human-like voice agents without coding.

BlockBeats reported on July 1st that xAI launched its new product, Voice Agent Builder. This no-code platform, based on Grok Voice, allows developers and operators to quickly build human-like conversational voice agents. It is now available for use, priced at $0.05 per minute. This tool targets users who want to directly deploy production-grade voice AI applications, lowering the technical barrier to building a voice system from scratch. The official statement indicates that the platform integrates complete capabilities including telephone system access, knowledge retrieval, tool invocation, content security control, and observability. It also supports integration with existing phone numbers, APIs, and MCPs from external systems. Compared to traditional voice technologies that require separate calls to speech-to-text, large language models, and text-to-speech services, leading to increased latency and costs, Voice Agent Builder achieves a unified voice processing workflow through deep coupling with the Grok model, reducing latency, minimizing points of failure, and improving overall performance. Click the original link below to join the Beating · Lark AI news channel for 24/7 monitoring of global AI hotspots and news.