On September 23, 2026, OpenAI once again surprised the tech industry. On the company's official YouTube channel, a barely minute-long video debuted, announcing fundamental changes to the ChatGPT Voice ecosystem. This concerns three pillars: voice integration with external tools (email, calendar, Slack), the debut of a new generation of GPT-6 models (under the code names Astra, Sol and Luna), and a completely new service, ChatGPT Work, designed to automate office tasks directly in the browser.
OpenAI is increasingly shifting its communication strategy. Instead of detailed, technical blog posts, the AI giant is betting on dynamic video materials. The latest clip, published on September 23 at 10:11 AM Pacific time, runs under a minute and bears the telling title Get Stuff Done with chatgpt Voice. Although sparse in form, its description hides announcements that could redefine how we interact with voice assistants.
Official OpenAI declarations: Three pillars of the new update
Before we move on to deeper analysis and verification of these reports, let's precisely examine the information OpenAI placed directly in the description under the published video. At this moment, this is the sole official source of knowledge. According to the creators' declarations, the upcoming novelties include:
- Voice plugins: “Use plugins such as email, calendar and Slack.” Users will gain the ability to directly manage these integrations via natural voice commands.
- Generation of GPT-6 models: “Powered by GPT-6 Astra, Sol and Luna.” This is the first official disclosure of the code names for the next generation of models, which have not appeared in any press releases so far.
- ChatGPT Work: “Used in chatgpt Work in web and mobile versions to create documents, presentations, websites and spreadsheets or perform complex tasks in the browser, simply by speaking.” This heralds an advanced enterprise‑class work environment.
- Global rollout: “Being deployed globally today in the latest version of the app.” The wording “today” refers directly to the publication date of the video, i.e., September 23.
All of the above declarations come directly from the description of the material on the official channel @openai. Although they sound revolutionary, as professional market observers we must note that they currently lack independent technical confirmation.
Verification status: Where is the evidence? (As of September 25, 2026)
As a tech editorial team, we conducted a detailed audit of available information channels and technical documentation. We analyzed the official OpenAI blog (openai.com/blog), API documentation (platform.openai.com/docs), the official press archive (openai.com/press/), the company's profile on platform X, and also the service status monitoring system (status.openai.com) as well as changelogs in the App Store and Google Play shops for September 23–24. Additionally, we checked leading tech media such as The Verge, TechCrunch and Ars Technica for publications concerning “chatgpt Voice” or “GPT-6” from September 2026.
The conclusions from this investigation are surprising: none of the key announcements have been externally confirmed so far.
- The API documentation contains no mention of voice support for plugins for email, calendar or Slack.
- OpenAI blog posts published after June 2024 are silent on the names “Astra”, “Sol”, “Luna” and “GPT-6” itself.
- OpenAI's official product offering lacks any information about a dedicated “chatgpt Work” service.
- Changelogs in the mobile app stores do not record the deployment of these features around the film's premiere date.
- Independent tech sites have not yet published any tests or confirmations of the functioning of the described novelties.
We are thus dealing with an atypical situation: OpenAI's official declarations from the video description have not yet found reflection in the technical infrastructure nor in independent experts' reports. This does not, of course, mean that the announced features are fictional. It may result from a gradual, phased rollout, delays in documentation updates, or a new marketing strategy in which this short video builds tension ahead of the proper debut.
What does this mean in practice? Analysis for users and business
Individual users and ChatGPT Voice
For individuals using ChatGPT Voice on a daily basis, the first step should be to verify the app version on iOS and Android systems. It is worth regularly checking the “What’s new” section as well as directly in the app settings – OpenAI is known for rolling out features in waves, meaning that novelties may appear for individual users at different times.
Enterprise sector and ChatGPT Work
For technology leaders and IT administrators, the announcement of ChatGPT Work sounds like a direct evolution of the ChatGPT Enterprise plan, with particular emphasis on the voice interface and autonomous creation of resources (documents, presentations, spreadsheets or pages). When planning an AI deployment strategy in an organization, it is worth closely monitoring the official channels for business customers. A helpful guide to understanding the current administrative framework can be a practical guide to managing chatgpt Work in large organizations – although it is based on the state of knowledge from July, it serves as an excellent reference point for the upcoming changes.
Creators and freelancers
The vision of generating complete presentations, websites or spreadsheets solely via voice commands represents a huge leap forward. Until now, ChatGPT could generate raw HTML, Markdown or CSV structures, but direct creation and publishing of ready‑made files (e.g., .PPTX, .XLSX) is a completely new level of integration. To better understand this evolution, it is worth reading a reminder of how plugins worked in 2024 and what remains of them in 2026. This will allow us to compare the old, text‑based plugin interfaces with the announced voice revolution.
Key question marks: What we still don't know?
Analyzing the dynamic development of artificial intelligence systems, we must pose several fundamental questions to which neither the short film nor its description provides answers:
- Are GPT-6 Astra, Sol and Luna three variants of the same model (e.g., optimized for speed, cost and performance), or completely different architectures?
- How will the licensing model for ChatGPT Work be structured? Will it be an entirely new pricing plan, or merely an extension of the existing Enterprise subscriptions?
- What will the configuration of voice plugins look like? Will integration with email, calendar and Slack require manual account pairing, akin to the solutions from 2024?
- How does OpenAI intend to guarantee the security and privacy of corporate data while performing complex operations in the browser via voice?
- Will the new features reach the European market without delays? One should remember that new EU guidelines on AI transparency impose strict informational obligations on technology providers.
Answers to these questions should appear in the coming weeks. We recommend regularly monitoring the community section on the OpenAI YouTube channel as well as their official profile on platform X.
Editorial comment: Evolution toward agents and a shift in PR strategy
The observed acceleration of OpenAI's release cycle and the turn toward video communication are a clear sign of the times. Short video formats generate enormous reach, although unfortunately this comes at the cost of technical precision. The code names “Astra”, “Sol” and “Luna” most likely represent varied gradations of the new GPT-6 model (analogous to the earlier split into standard, Turbo and “o” versions). Meanwhile, “ChatGPT Work” appears as a logical step toward fully autonomous AI agents capable of operating directly in the user's browser. The whole picture fits perfectly into the market trend where voice becomes the primary interface for controlling complex digital tools.
Regardless of enthusiasm, as of September 25, 2026 we must rely solely on the textual declarations beneath the video. Before making any strategic business or investment decisions, we recommend holding off until official technical documentation appears.
Recommended steps: How to prepare?
- Update the ChatGPT app on your mobile device and verify the availability of new options in the Voice settings as well as the presence of the “Work” tab.
- Monitor the official OpenAI blog and API documentation, where detailed technical specifications of the announced models and plugins should appear.
- Keep track of the operational status of services on the status.openai.com page for planned maintenance work or phased rollouts of new functionalities.
- If your organization uses the ChatGPT Enterprise plan, contact your dedicated OpenAI customer success manager directly to confirm the status of ChatGPT Work and migration conditions.
We will closely monitor this topic. As soon as hard, technical evidence or official confirmation from OpenAI appears, we will immediately update this article. For now, treat the published video as a declaration of development direction, not a ready‑to‑deploy list of changes in production code.
Primary source: The video “Get Stuff Done with chatgpt Voice” on the OpenAI YouTube channel, published September 23, 2026, video description. External verification: no confirmation in OpenAI sources (blog, API, press, status) nor in tech media up to September 25, 2026.
Comments