OpenAI’s DevDay 2026 was all about turning ChatGPT into a larger platform for agents, developers, and teams.
The company unveiled more than 20 announcements spanning always-on AI agents, new models, coding tools, enterprise security, plug-ins, and collaborative workspaces.
The biggest launch was Dots, OpenAI’s new always-on agents that can work across applications and continue pursuing tasks in the background. The company also introduced GPT-6.1 Sol, a cheaper model with near-GPT-6 Astra capabilities, along with a new Ultrafast tier for users who prioritise speed.
For developers, OpenAI brought Codex to the cloud, added computer-use capabilities to its Agents API, and launched a Decisions API for structured decision-making.
Here are all the major announcements from OpenAI DevDay 2026:
1. Dots
OpenAI describes Dots as always-on agents that can take responsibility for ongoing work rather than waiting for users to send individual prompts.
Powered by GPT-6 Astra, Dots can operate on its own cloud computers, use a web browser, and interact with more than 4,000 supported apps. Users can communicate with a Dot through text or voice, while integrations with Slack and Microsoft Teams allow it to operate inside workplace conversations.
Dots can also use OpenAI’s Codex and ChatGPT Work capabilities to research information, analyse data, create documents, and build software. The idea is that a Dot can continue working on a project as circumstances change instead of completing a single isolated task.
Dots is available on Pro and Business Premium in eligible markets. Enterprise, Edu, and Healthcare customers can access a beta when administrators enable it.
2. GPT-6.1 Sol
OpenAI introduced GPT-6.1 Sol as an upgrade to GPT-6 Sol, targeting developers and professionals who need strong reasoning and agentic capabilities without paying Astra-level prices.
The company says GPT-6.1 Sol is particularly strong at agentic coding, computer use, and professional work. OpenAI positions it close to GPT-6 Astra in capability while pricing it at one-fifth of Astra’s standard input and output token prices.
That makes the model particularly relevant for applications that need to run many AI tasks. Lower token costs can allow developers to execute longer workflows, run more agents, or increase usage within the same budget.
GPT-6.1 Sol is available through the API and across Plus, Pro, Business, Enterprise, and Edu. OpenAI has also applied the same core safety safeguards used for GPT-6 Astra to the new model.
3. Private Intelligence
OpenAI introduced Private Intelligence as a collection of privacy and security capabilities for businesses using its models.
One component is Zero Data Retention with Private Safety Processing. It allows OpenAI to perform automated safety reviews without giving OpenAI personnel access to the underlying customer content.
The company is also previewing Private Inference, which is expected later this year. It combines confidential computing with controls that OpenAI says can be independently verified.
4. Codex Moves Into the Cloud
OpenAI is bringing Codex to the cloud with Codex in the cloud. Developers can now start and continue coding tasks from computers, phones and other devices while Codex runs work remotely.
Each task can use reusable development environments containing the repositories, tools, dependencies and permissions needed for the project.
5. Sign-in With ChatGPT
OpenAI will now allow users to connect their ChatGPT account with external tools and use their existing subscription benefits in participating services.
The company says users can apply their ChatGPT allowance across 16 partners, including Cognition’s Devin, Notion, Vercel, T3, OpenClaw, and Dactyl. They can also control how much of their plan allowance each participating tool can consume.
The feature combines two functions. Users can sign in to supported tools using their ChatGPT account, while eligible Plus and Pro subscribers can also use their OpenAI usage allowance in participating applications.
OpenAI says identity support is available globally, while plan-based usage is currently limited to Plus and Pro users in participating tools.
6. $500 Pro Tier
OpenAI introduced a new $500-per-month Pro plan for users who need the highest usage limits and faster model performance.
The plan provides 25 times the usage allowance of ChatGPT Plus and includes access to the new Ultrafast tier. It is targeted at users running demanding workloads across OpenAI’s models, Codex, and ChatGPT Work.
The plan also provides access to GPT-6 Astra Ultrafast, which OpenAI says can generate tokens substantially faster than standard inference.
7. OpenAI Marketplace
OpenAI announced an OpenAI Marketplace in beta for eligible enterprise customers. The programme allows companies to apply part of their existing OpenAI commitment towards approved partner software.
The first group includes 32 partners spanning several categories. These include Figma for creative work; Adobe, Sierra, Decagon, HubSpot, Salesforce, and ServiceNow for customer experience; Harvey and Legora for legal work; Palo Alto Networks and CrowdStrike for cybersecurity; and Baseten for open-source models.
The marketplace gives enterprises more flexibility in how they use their existing AI budgets.
8. ChatGPT Space
ChatGPT Space is a new collaborative environment combining ChatGPT and Dots.
Teams can create a dedicated Space for a project, provide shared knowledge, and give ChatGPT instructions about how the workspace should be organised.
The goal is to make it easier for teams to continue work without repeatedly reconstructing context. ChatGPT can organise information within the Space so users can find previous work and resume projects.
The feature is available to Pro, Business, and Enterprise users on desktop and web, while finding, reading and sharing Pages is also supported on mobile.
9. Agents API with Computer Use
OpenAI has added computer-use capabilities to Agents API, allowing developers to build agents that interact directly with software. The agents can operate applications and complete tasks through a computer interface rather than relying only on APIs or predefined integrations.
OpenAI is also bringing several capabilities from Codex into Agents API, including multi-agent workflows, tool search, tool calling, and context compaction.
The capability is available through the API and also powers experiences in Codex and ChatGPT Work for eligible users.
10. ChatGPT Comes to Slack, Microsoft Teams
OpenAI is bringing ChatGPT directly into workplace conversations through Slack and Microsoft Teams.
Employees can mention ‘@ChatGPT’ in a channel, thread, or direct message and ask it to help with work. The assistant can use tools connected by administrators or tools available through the user’s own permissions.
OpenAI says employees do not need an individual ChatGPT licence to participate in these shared interactions on Business and Enterprise plans.
11. Meetings Plug-in
OpenAI introduced a Meetings plugin that can record meeting notes and save personalised summaries and action items in ChatGPT Space.
Users can keep those notes private or share them with their team. Users can share the information for follow-up work, such as updating a project plan or drafting a message.
OpenAI says audio is deleted once the notes are prepared and cannot subsequently be accessed or replayed.
The plug-in is available in beta through the ChatGPT desktop app on macOS for Pro and Business users, with Enterprise support planned.
12. Plugin Extensions
OpenAI is opening the platform for developers to create their own interactive experiences inside ChatGPT.
Plugin Extensions allow developers to give their plug-ins a dedicated home in the ChatGPT sidebar. They can also build interactive panels that sit alongside conversations and create viewers for supported file types.
OpenAI says the capability is available across all plans.
Your reaction
Discussion
What's your take on this story?
No comments yet — be the first to share your view.