DevDay keynote time was set for 10am PT on Tuesday 29 September, with Sam Altman on stage at Fort Mason in San Francisco and a livestream for everyone else. The night before, he posted a characteristically cryptic teaser: “Pretty excited for DevDay tomorrow. We have found a new thing.” By the time he left the stage, OpenAI had shown more than 20 products and features to roughly 2,500 people in the room.
The new thing turned out to be Dots, always-on AI agents that OpenAI is pitching against Meta’s Muse. Around it came a new model, GPT-6.1 Sol, a faster tier called Ultrafast, a $500-a-month Pro plan, a way for developers to build full applications inside ChatGPT, and a privacy offer for large customers. OpenAI also said ChatGPT now has 1.2 billion weekly users.
This article covers when and where the DevDay keynote ran, what each major launch does and who can use it, how the day compared with the leaks and reports that preceded it, the numbers OpenAI shared, the safety debate hanging over the event, and what developers and businesses should do next.
Table of contents
- When and Where the DevDay Keynote Took Place
- Dots Led the DevDay Keynote
- GPT-6.1 Sol, Ultrafast and Pro 500
- Everything Else Announced at the DevDay Keynote
- What OpenAI Teased Before the DevDay Keynote and What Shipped
- The Numbers From DevDay
- The Safety Backdrop to the DevDay Keynote
- What Developers and Businesses Should Do Next
- DevDay Keynote FAQ
- References
When and Where the DevDay Keynote Took Place
OpenAI billed DevDay 2026 as its largest developer conference to date. In-person places were by application, and those applications had already closed by the day of the event. Everyone else could watch the DevDay keynote and sessions free online, with recordings published afterwards.
The day’s schedule
Crypto Briefing published the run of show on the morning of the event. The DevDay keynote sat in the 10am slot, with breakout sessions filling the afternoon.
| Time (PT) | Time (BST) | Session |
|---|---|---|
| 8:00am | 4:00pm | Breakfast for in-person attendees |
| 10:00am | 6:00pm | Keynote with Sam Altman, livestreamed |
| 11:15am to 3:30pm | 7:15pm to 11:30pm | Breakout sessions |
| 4:00pm | Midnight | Closing session |
| 4:45pm to 7:00pm | 12:45am to 3:00am | Reception |
For UK readers, that put the DevDay keynote at 6pm British Summer Time, a friendlier slot than many US launches. OpenAI also said it would follow the event with smaller “DevDay Exchanges” in cities around the world.
Who took the stage
Altman led the DevDay keynote. Crypto Briefing named Romain Huet, Holly Li and Tejal Patwardhan as the other speakers joining him on stage.
The tease the night before
Altman’s “we have found a new thing” post set expectations for something beyond the usual model update. Code references had already pointed to an always-on agent, reported under the names “o” and later Aeon, and our coverage of OpenAI’s always-on assistant tracked those leaks. The DevDay keynote confirmed the idea under a new brand.
Dots Led the DevDay Keynote
Dots were the centrepiece. “Dots are remarkably capable, always-on agents built to handle really anything you can think of,” Altman said during the DevDay keynote, according to CNBC. “They bring AI into a whole new form factor to work and give you more time and attention so you can do the things that you care about.”
What a dot does
OpenAI’s launch post says Dots are powered by GPT-6 Astra, have their own cloud computer and browser, learn from feedback over time and can work towards your goals around the clock. Through OpenAI’s plugins they can connect to more than 4,000 apps. You can message or call a dot in ChatGPT on desktop, web and mobile, and message it in Slack and Microsoft Teams, with texting “coming soon”.
Altman made the pitch personal. “I’ve been really surprised at what a difference this makes to my life,” he said during the DevDay keynote, according to CNBC. “I feel like I’ve finally gotten some of my attention back. I no longer feel quite as addicted to my phone in the same way.”
Users start with a primary dot, give it a name and customise it. OpenAI says it envisages teams of dots working together later, and it previewed “specialist dots” with their own identity and credentials for enterprise jobs such as procurement, invoice processing and customer support. It is working with Microsoft to bring those under Agent 365 governance controls.
Who gets it
Dots are rolling out to ChatGPT Pro and Business Premium users in eligible markets. Enterprise, Edu and Healthcare workspaces can try a beta, which is off by default until an administrator enables it. The first dot is included in those plans at no extra cost, and conversations with it do not count towards ChatGPT usage limits, though tasks it starts in Codex or ChatGPT Work do. Our earlier reports on the switch from Aeons to Dots followed the branding change.
How permissions work
OpenAI’s launch post spends a lot of space on control, and with reason: the same week, Meta’s Muse agent shared a YouTuber’s home address with Marketplace buyers. OpenAI says dots start with built-in rules for when to act and when to ask, that Custom Rules let users allow, require approval for or block specific actions, and that an auto-review step checks actions that could affect accounts or “share information”. Background “proactive research” uses read-only tools, and “certain sensitive tasks, such as changing a password, always stay with you”.
GPT-6.1 Sol, Ultrafast and Pro 500
The second block of the DevDay keynote covered models, speed and pricing, the announcements most developers and heavy users will feel first.
GPT-6.1 Sol arrives a week after GPT-6 Sol
OpenAI introduced GPT-6.1 Sol one week after rolling out GPT-6 Sol, calling it a “major upgrade” with strong performance on professional work, computer use and agentic coding. CFO Sarah Friar later highlighted a price drop for the model in an interview with CNBC. The release matters partly for what it replaced: a day earlier, OpenAI had pulled plans to launch GPT-6.1 Astra because it did not meet the company’s safety standards, a story we covered in OpenAI’s cancellation of GPT-6.1 Astra.
Ultrafast speeds
Ultrafast is a premium speed tier across ChatGPT, Codex and the API. OpenAI said it generates tokens up to eight times faster in Codex and up to six times faster in the API. “We think you are really going to love this and not want to go back,” Altman said. It follows the Ultrafast API expansion reported the previous week.
A $500 plan and a reopened $200 plan
Altman announced Pro 500, a $500-a-month ChatGPT plan with the company’s highest usage limits and access to Ultrafast in ChatGPT, ChatGPT Work and Codex. At the same time OpenAI reopened sign-ups to its $200 Pro plan, which it had paused after demand for GPT-6 Astra strained its systems. OpenAI’s Thibault Sottiaux said the plan’s usage would be calculated differently, netting out “at half the dollar in API spend compared to the old Pro $200 plan”.
Friar told CNBC that when OpenAI launched the $200 plan, “people thought we’d lost our minds”, but that customers are now moving towards consumption-based spending rather than only per-seat pricing.
Everything Else Announced at the DevDay Keynote
OpenAI’s recap post lists more than 20 announcements. The table groups the main ones with the availability OpenAI gave.
| Announcement | What it is | Availability stated |
|---|---|---|
| Dots | Always-on agents with their own cloud computer | Pro and Business Premium; Enterprise, Edu, Healthcare beta |
| GPT-6.1 Sol | New model for professional work, computer use and coding | Announced at the event |
| Ultrafast | Premium speed tier | ChatGPT, Codex and the API |
| Pro 500 | $500-a-month plan with the highest limits | Consumer plan |
| Private Intelligence (preview) | Zero Data Retention with private safety processing, plus private inference | Preview for large customers |
| Codex anywhere | Run Codex on a computer, from a phone or in the cloud, with reusable environments | Plus, Pro, Business, Healthcare, Education, Enterprise |
| Plugin extensions | Full apps inside ChatGPT and Codex: sidebar panels and file viewers | All plans |
| ChatGPT Space and Pages | Shared team space and documents for people and agents | Pro, Business and Enterprise on desktop and web |
| ChatGPT in Slack and Teams | @ChatGPT in team chats, plus a meetings plugin | Shown in the recap |
| Sign in with ChatGPT and plan usage | Use a ChatGPT login and plan allowance in 16 partner tools | Identity global; plan usage for Plus and Pro |
| Developer platform updates | Decisions API, computer use in the Agents API, MCP events for plugin automations | Shown in the recap |
For developers building on ChatGPT
The developer story at the DevDay keynote was about ChatGPT as a platform. “With plugin extensions, you can build an editor, a dashboard, a whole workspace directly into ChatGPT and Codex,” Altman said, showing a meetings app that displayed a user’s upcoming meetings inside the chatbot. OpenAI also showed improved plugin creation, submission and discovery, Sites with connected plugins, and MCP events that trigger plugin automations.
Sign in with ChatGPT lets people log in to other tools with their ChatGPT account. More importantly, Plus and Pro users can spend their plan allowance in 16 partner tools, including Cognition’s Devin, Notion, Vercel, T3, OpenClaw and Dactyl, and control how much each can use. As developer Hamel Husain summarised it on X during the DevDay keynote, it “lets people sign into your app and use the tokens they are paying for already”.
For software teams using Codex
Codex got its own block of the DevDay keynote. OpenAI says developers can now run Codex wherever they need it: on a computer, remotely from a phone, or in the cloud from any device. Reusable development environments let tasks start quickly and give a team a shared setup with approved settings and permissions. The recap also shows an updated Codex command-line tool, a code review view and Codex Security Cloud, which collects security findings in one dashboard.
The cloud option is available on Plus, Pro, Business, Healthcare, Education and Enterprise plans. For teams already using Codex locally, the practical change is that long-running jobs no longer need a laptop left open, and a lead can define one environment for everyone rather than debugging each developer’s setup.
For teams
ChatGPT Space is a shared home where teammates, ChatGPT and a dot build on the same knowledge, and Pages is a new document type for people and agents to write, create images and build charts together. Collaborative presentations, shared team tasks, shareable profiles and a meetings plugin round out the workplace push. We covered the early signs in our report on OpenAI Spaces for human and agent collaboration.
For privacy-sensitive customers
Altman previewed OpenAI Private Intelligence, which he said “delivers frontier model safety without storing your content on our servers at all”, with private inference providing privacy at the moment the model runs. OpenAI said it designed the systems with large customers including Cisco, Databricks and Snowflake. For security teams, Codex Security Cloud adds a cybersecurity findings dashboard for code, which fits the same enterprise pitch.
What OpenAI Teased Before the DevDay Keynote and What Shipped
Much of the DevDay keynote had been previewed through leaks, code strings and reports in the weeks before. Comparing the two is a useful check on which reports were right.
| What was reported before | What the DevDay keynote delivered |
|---|---|
| An always-on assistant called “o”, then Aeon | Dots, always-on agents for Pro and Business Premium |
| Managed agents and environments for DevDay | Codex cloud environments and computer use in the Agents API |
| Spaces for human and agent collaboration | ChatGPT Space plus Pages |
| A $500 Pro tier | Pro 500 at $500 a month, with Ultrafast |
| Wider Ultrafast API access | Ultrafast in ChatGPT, Codex and the API |
| GPT-6.1 Astra | Withheld on safety grounds; GPT-6.1 Sol launched instead |
| Pro sign-ups paused over Astra demand | $200 Pro plan reopened with new usage limits |
The pattern is clear: nearly every major leak landed, usually under a different name. The biggest change from expectations was not a product but a model left out, as Astra’s withdrawal shaped how the whole DevDay keynote was framed.
The Numbers From DevDay
OpenAI and its executives shared several figures during and around the DevDay keynote. The recap puts ChatGPT at 1.2 billion weekly users, up from the 1 billion milestone reported in July.
Friar told CNBC that OpenAI had an “incredible” third quarter, confirming 70% quarter-on-quarter growth and saying its enterprise business has doubled since July. She declined to discuss reported early talks with investors about a new funding round, which Bloomberg first reported at around $30 billion, noting that OpenAI raised $122 billion in March at an $852 billion valuation and is “very well capitalized”.
Scale of the event itself
CNBC counted more than 20 products and features shown to roughly 2,500 attendees. Altman closed the DevDay keynote on a broad note: “AI should be about giving people more power over their own lives, more tools to create, learn, discover, expand knowledge,” he said, adding that OpenAI hoped its launches would help get there. OpenAI’s recap describes “more than 20 major announcements across ChatGPT, Codex, our models, and entirely new forms of working with AI”.
The Safety Backdrop to the DevDay Keynote
The DevDay keynote came at a tense moment. The Verge noted that the industry had been shaken by agents hacking outside companies, including OpenAI’s own models breaching Hugging Face earlier in the year, which fuelled calls to slow AI development. OpenAI had endorsed a call to slow the pace of advanced model development earlier in September, and our report on the five ways rogue AI agents are affecting the internet covered OpenAI’s own analysis.
Holding back Astra
Altman used the day to explain why GPT-6.1 Astra was withheld, and said OpenAI is doing something similar to Nvidia’s agent safety platform. Friar put it simply: “When we have to pace the frontier, we’ll do that. That’s what we’re showing right now.” She added that there will also be times when OpenAI continues to “show model progression”.
Protests and an IPO question
Protesters gathered outside Fort Mason before the DevDay keynote, with signs reading “Drop the ICE contract” and “People over AI”, according to The Verge. More than a dozen organisations sponsored the rally. Separately, Altman told CNBC he has no particular timeline for a stock market listing and does not want to be “barreling forwards with an IPO” while “things feel so in flux”.
What Developers and Businesses Should Do Next
The DevDay keynote announced a lot, and not all of it is available everywhere yet. A few practical steps make sense now.
For developers
Read the plugin extensions documentation before building a separate web app for ChatGPT users; a native panel may reach more people with less work. If your product has its own AI costs, look at Sign in with ChatGPT and plan usage, which could let users bring their existing allowance. Test Ultrafast on latency-sensitive paths, but budget for it as a premium tier rather than a default.
For businesses
Treat Dots as you would any new staff member with system access: decide what it may do alone, what needs approval and what stays with people. Enterprise administrators control whether the beta is switched on, which gives time to write those rules first. If data residency or retention has kept you away from frontier models, ask about the Private Intelligence preview. For an overview of the models and tools behind these launches, see our AI models, tools and releases hub.
DevDay Keynote FAQ
When was the OpenAI DevDay keynote?
The DevDay keynote started at 10am PT (6pm BST) on Tuesday 29 September 2026 at Fort Mason in San Francisco, and was livestreamed free with recordings available afterwards.
What was the biggest announcement?
Dots, always-on AI agents powered by GPT-6 Astra that have their own cloud computer, connect to more than 4,000 apps and can be messaged in ChatGPT, Slack and Teams. They are rolling out to Pro and Business Premium users first.
What is Pro 500?
Pro 500 is a new $500-a-month ChatGPT plan with OpenAI’s highest usage limits and access to the Ultrafast speed tier in ChatGPT, ChatGPT Work and Codex. The $200 Pro plan also reopened with new usage limits.
How fast is Ultrafast?
OpenAI says Ultrafast generates tokens up to eight times faster in Codex and up to six times faster in the API.
How many people use ChatGPT?
OpenAI’s DevDay recap says ChatGPT has 1.2 billion weekly users, up from 1 billion in July.
References
OpenAI sets DevDay keynote for 10am PT, promises new features (Crypto Briefing)
OpenAI DevDay 2026: Live updates and announcements (CNBC)
OpenAI DevDay 2026: The biggest news and announcements (The Verge)
OpenAI is adding a $500 a month ChatGPT Pro tier (The Verge)
More AI coverage: explore Progressive Robot's AI Models, Tools & Releases hub — hands-on reviews, setup guides and benchmarks in one place.
