MiniMax Unveils Major Overhaul with Code 2.0 Desktop Update Featuring Enhanced Architecture

Stock News
07/16

MiniMax has officially launched the updated desktop version of its MiniMax Code 2.0 platform. This release is built upon the open-source Pi Agent framework and features a comprehensive reconstruction of the underlying architecture, focusing on improving session start-up speed, stability during long-duration tasks, and the contextual linking capabilities for tool invocation.

The new version also includes optimizations for features like file previews and chart interactions, with plans to further expand its financial research capabilities. With the revamped architecture, MiniMax has rebuilt the pathways for session operation, state management, and tool calls. This addresses issues such as waiting periods, interruptions, and context loss that can occur during extended tasks.

From a user experience perspective, the wait time from creating a new session to receiving the first output has been significantly reduced. The frequency of abnormal interruptions and silent freezes during conversations has also been substantially decreased. For complex tasks that involve continuously reading materials, calling tools, modifying files, and repeated verification, the new version improves the mechanisms for saving and transferring session states and tool-call information. This lowers the risk of accidental session loss and execution context breaks, enabling the Agent to maintain context across multiple task stages.

Beyond the core architectural upgrades, MiniMax Code has refined high-frequency functions like file previews and charts. New capabilities include support for editing within a selected area of the file preview window, along with enhanced chart loading stability. These improvements are designed to facilitate report creation, presentations, and subsequent analysis.

In terms of professional applications, MiniMax is expanding the Agent's capabilities in financial research, investment analysis, and risk management. The company has already integrated with applications such as the Hang Seng Financial Database and Qichacha MCP, connecting global financial market data and domestic market entity dynamics into its financial module. This module is currently available for trial on the web version and is scheduled for release on the MiniMax Code desktop client in the near future.

By connecting to financial databases, MiniMax Code can perform information retrieval, data organization, indicator comparisons, and report generation based on user-defined research objectives. This reduces the need for users to repeatedly query, download, and consolidate information across different databases and platforms. For instance, when researching the A-share storage industry chain, MiniMax Code can outline various segments including storage chips, memory modules, controller chips, packaging & testing, storage devices, and upstream materials and equipment. It can distinguish between companies for which this is a core business and those with peripheral involvement. The platform can then compare metrics such as stock price performance, valuation changes, capital flows, institutional research interest, and consensus forecast revisions, ultimately generating a structured industry research report.

It is reported that MiniMax Code will launch additional features this month, including remote control, browser manipulation, target mode, and planning mode, further extending its ability to handle complex tasks and facilitate cross-environment collaboration. During the WAIC 2026 conference, attendees can visit the MiniMax booth (H1-A801) at the Shanghai World Expo Exhibition & Convention Center to experience immersive interaction with the company's full-modal models and latest products firsthand.

免責聲明:投資有風險,本文並非投資建議,以上內容不應被視為任何金融產品的購買或出售要約、建議或邀請,作者或其他用戶的任何相關討論、評論或帖子也不應被視為此類內容。本文僅供一般參考,不考慮您的個人投資目標、財務狀況或需求。TTM對信息的準確性和完整性不承擔任何責任或保證,投資者應自行研究並在投資前尋求專業建議。

熱議股票

  1. 1
     
     
     
     
  2. 2
     
     
     
     
  3. 3
     
     
     
     
  4. 4
     
     
     
     
  5. 5
     
     
     
     
  6. 6
     
     
     
     
  7. 7
     
     
     
     
  8. 8
     
     
     
     
  9. 9
     
     
     
     
  10. 10