OpenAI Altmanan Declare ChatGPT Final OS entrance: 8 minutes for smarts and 800 million for active users

On October 7th, the technological media, The Decoder, published an article yesterday, October 6, reporting that OpenAI On Development Day, Sam Altman, the chief executive officer of the company, announced a number of major updatespromote ChatGPT APPLY AND AI FROM A CHAT TOOL TO SET AgentTransformation of integrated and integrated platforms。

OpenAI Altmanan Declare ChatGPT Final OS entrance: 8 minutes for smarts and 800 million for active users

1AI QUOTES BOVEN, WHERE THE EVENT LASTED ABOUT 50 MINUTES, AND ALTMANN PRESENTS FOUR NEW DEVELOPERS ' TOOLS, AS FOLLOWS:

  • Apps SDK (Application Development Toolkit): a set of tools provided by OpenAI to enable software developers to create a powerful embedded application for ChatGPT, like the development of the App. Users can use these applications directly in chats without having to jump to other websites。
  • AgentKit (intelligence development package): Another set of tools launched by OpenAI dedicated to the creation of "AI Intelligence". This intelligent body can understand user needs and perform a complex set of tasks as a smart assistant, and the visual interface provided by AgentKit makes the creation process as simple as building blocks。
  • Codex Programming Assistant officially online: The new version enhances teamwork by adding Slack integration and SDK for custom workflows。
  • EXPAND APIGPT-5 Pro, gpt-realtime-mini and Sora 2 have now joined API。

i. Apps SDK: ChatGPT platform, primary application replacement plugin

OpenAI launched the new Apps SDK (Applications Development Toolkit) at this event, which aims to promote the transformation of ChatGPT from a chat tool to a mature third-party application platform。

Using the SDK, the developer can directly build a fully functional interactive application within ChatGPT to support fullscreen mode, user login and payment protocols. This means that without leaving the chat interface, users can call Canva to design posters, search for properties through Zillow or play music using Spotify。

The first online applications include Booking.com, Canva, Expedia, etc., presage a more integrated function and experience of a more seamless primary application to replace a single one of the previous functions。

II. AgentKit: Visualizing tools to enable everyone to build AI smart bodies

OpenAI has launched the tool kit for building an autonomous AI smart body. At its core, Agent Builder, a visual dragr, allows developers to combine workflows like a flowchart and quickly create, deploy and optimize intelligent bodies capable of carrying out complex tasks on their own。

In a live presentation, a smart body that recommends a programme of meetings was built and published in only eight minutes. The package also contains components such as Evals for testing and Contractor Registry for managing data access, and contains a security fence designed to significantly lower the development threshold for AI intelligence。

iii. Codex is officially online and user data is highly innovative

During this event, OpenAI announced that its programming assistant, Codex, was on line (GA). Since May, when cloud intelligence was introduced, the use of Codex has increased tenfold. The new version enhances teamwork by adding Slack integration and SDK for custom workflows。

OpenAI has published the latest operating data: ChatGPT ' s weekly active users have exceeded 800 million, platform developers have exceeded 4 million, and API traffic has reached an alarming level of 6 billion Token per minute。

IV. API model matrix update: GPT-5 Pro and Sora 2

OpenAI has opened the API interface of its most advanced language model GPT-5 Pro, which is designed to handle complex tasks, supports text and image input, and has context windows up to 400,000 Token. At the same time, the popular Vincent Sora 2 video model has been made available through API。

In order to meet the needs of different scenarios, OpenAI also introduced two high-value models: gpt-realtime-mini, about 70% less than previous real-time API; and gpt-image-1-mini, which reduced image processing costs by about 80%, which helps developers to scale up applications at lower cost。

statement:The content of the source of public various media platforms, if the inclusion of the content violates your rights and interests, please contact the mailbox, this site will be the first time to deal with.
Information

OpenAI adjusts the copyright policy, Sora will allow the active animators to come in

2025-10-7 12:06:35

Information

Google "Nano Banana" image AI model is officially open for commercial use: synchronise new batch processing, customize long-width ratio

2025-10-7 13:44:27

Search