OpenAI DevDay 2026 took place on September 29, 2026 in San Francisco, and it was the company's biggest developer event so far. Sam Altman, Romain Huet, Tejal Patwardhan and Holly Li walked through more than 20 launches spanning ChatGPT, Codex, the API and a brand-new kind of AI agent called dots.
If you only have two minutes, the headline is simple: OpenAI wants AI that works in the background on your behalf, not just a chatbot that answers questions. Below is every major OpenAI DevDay 2026 announcement, what it costs, who gets it, and what it means if you build for the web.
OpenAI DevDay 2026 at a glance
- Dots: always-on AI agents powered by GPT-6 Astra, each with its own cloud computer and browser.
- GPT-6.1 Sol: near-Astra performance at one-fifth of Astra's API price ($2 input / $10 output per million tokens).
- Ultrafast: a premium speed tier with up to 8x faster generation in Codex and up to 6x in the API.
- Codex in the cloud: start and steer coding tasks from a laptop, a phone or any browser.
- Agents API with computer use: build agents that operate software, with OpenAI running the infrastructure.
- Decisions API: fast, constrained answers for classification and routing.
- Plugin extensions and MCP Events: developers can build native experiences inside ChatGPT.
- ChatGPT Space and Pages: shared workspaces where teammates and agents collaborate.
- Pro 500: a new top subscription tier with 25x Plus usage and Ultrafast access.
- Private Intelligence: stronger data protections for businesses using frontier models.
Dots: OpenAI's always-on AI agents
The star of OpenAI DevDay 2026 was dots. A dot is a persistent agent that learns your goals and keeps working toward them around the clock. Each one runs on GPT-6 Astra and has its own cloud computer and browser. Through ChatGPT's plugin ecosystem, a dot can connect to more than 4,000 apps (OpenAI).
You can reach your dot in ChatGPT on desktop, web and mobile, in Slack or Microsoft Teams, or on a voice call. Texting support is coming soon. Context follows the dot across every channel, so a project started in ChatGPT can continue in a Slack thread without a recap.
When you are not actively working with it, a dot runs "proactive research", scanning your connected apps with read-only tools to find useful next steps. Custom Rules let you allow, require approval for, or block specific actions. Sensitive tasks such as changing a password always stay with you.
Dots are rolling out to Pro and Business Premium subscribers in eligible markets, and Enterprise, Edu and Healthcare admins can switch on a beta. Your first dot is included in the plan, and chatting with it does not count against ChatGPT usage limits. BGR notes that dots are not included on the cheapest $100 Pro tier (BGR).
Dots arrive weeks after Meta's Muse and SpaceXAI's Grok Bot, so the always-on agent race is now fully underway. For a full breakdown, read our guide: OpenAI Dots Explained.
GPT-6.1 Sol: near-Astra intelligence at a fifth of the price
OpenAI also launched GPT-6.1 Sol, an upgrade to GPT-6 Sol aimed at agentic coding, computer use and professional work. OpenAI says it nearly matches GPT-6 Astra on those tasks while costing one-fifth as much per token (OpenAI).
| Model | Input ($/1M tokens) | Cached input ($/1M) | Output ($/1M) | Best for |
|---|---|---|---|---|
| GPT-6 Astra | $10.00 | $1.00 | $50.00 | Hardest reasoning and research tasks |
| GPT-6.1 Sol | $2.00 | $0.10 | $10.00 | Coding agents, computer use, business workflows |
| GPT-6 Luna | $0.10 | $0.01 | $0.50 | High-volume, everyday tasks |
The model ID is gpt-6.1-sol. In ChatGPT it is available in ChatGPT Work and Codex for Plus, Pro, Business, Enterprise and Edu users, but not yet in standard Chat. The cached input price of $0.10 is a big deal for agents that reuse long context across many calls.
We cover the benchmarks in detail in GPT-6.1 Sol: Pricing, Benchmarks and What It Means for Developers.
Ultrafast: a premium speed tier
Ultrafast is a new paid speed tier for latency-sensitive work. OpenAI quotes up to 8x faster token generation in Codex (around 300 tokens per second) and up to 6x in the API. GPT-6 Astra Ultrafast is live now in the API and in ChatGPT Work and Codex on Pro 500 and Enterprise plans (OpenAI).
Speed is not free: BGR reports Ultrafast costs six times the standard API price. A GPT-6.1 Sol Ultrafast version is coming soon.
Codex: coding from anywhere
Codex got four meaningful updates at DevDay 2026:
- Codex in the cloud: run tasks on your computer, remotely from your phone, or fully in the cloud, with reusable team environments. Available on Plus, Pro, Business, Healthcare, Education and Enterprise.
- Refreshed Codex CLI: voice control, a new
/agentsview for tracking parallel tasks, better session resume and worktree support. Available on all plans. - Code Review: a review experience in the ChatGPT desktop app with summaries, diffs and automatic first-pass reviews for GitHub pull requests and GitLab merge requests.
- Codex Security Cloud: scheduled or on-demand repository scans that triage findings and prepare fixes, even with your laptop closed. Available on Pro, Business, Enterprise and Edu.
During the keynote, Romain Huet had Codex build a raffle app live on stage that picked six attendees to win tickets to next year's event (Engadget).
New APIs for developers
- Agents API with computer use: agents can now operate software, and the API brings Codex's multi-agent orchestration, tool search, tool calling and context compaction into your own app. OpenAI hosts the infrastructure.
- Decisions API: points GPT-6 Luna at a fixed set of questions with predefined answers, useful for classifying content, routing requests or choosing an agent's next step. It is in limited preview, with broad release planned in the coming days.
- Bedrock Managed Agents, powered by OpenAI: built with Amazon so teams can run OpenAI agents entirely inside AWS.
Our developer deep dive, DevDay 2026 for Developers, covers these in more detail.
ChatGPT plugins become a real platform
OpenAI is opening the same platform it uses to build ChatGPT features. Plugin extensions let your plugin live in the ChatGPT sidebar, render interactive panels beside the conversation and add viewers for your file types.
A new Plugin Creator and redesigned submission flow should make publishing easier, and better ranking should help users find plugins. Plugins can now also run inside Sites, and support for the proposed MCP Events spec means plugins can trigger automations when something happens in a connected app.
Collaboration: Space, Pages, slides and @ChatGPT
- ChatGPT Space: a shared home where teammates, ChatGPT and dots build on shared knowledge (Pro, Business, Enterprise).
- Pages: a collaborative document type built for humans and agents to edit together.
- Collaborative slides: multi-user, agent-assisted decks with PowerPoint and Google Slides export, arriving in the coming weeks.
- Teams and shared tasks: recurring work, like weekly updates, delegated to scheduled or event-triggered team tasks.
- @ChatGPT in Slack and Microsoft Teams: mention ChatGPT in any channel or thread; teammates can join in without their own license.
- Meetings plugin: a macOS beta that takes notes, saves summaries to Space and deletes audio once notes are ready.
- Shareable profiles: showcase and share the Sites and plugins you build.
ChatGPT plan changes: Pro 500, Sign in with ChatGPT and Marketplace
The new Pro 500 plan offers 25x the ChatGPT Plus allowance and includes Ultrafast. According to OpenAI's Thibault Sottiaux, the $200 Pro plan now offers 10x Plus usage instead of 20x, though existing subscribers keep 20x for now, and the $100 tier stays at 5x (BGR).
Sign in with ChatGPT lets Plus and Pro users spend their plan allowance in 16 partner tools, including Cognition's Devin, Notion and Vercel. The OpenAI Marketplace lets eligible enterprises apply part of their OpenAI commitment to 32 approved partners such as Figma, Salesforce, Harvey and CrowdStrike.
Private Intelligence and the safety backdrop
Sam Altman previewed OpenAI Private Intelligence for businesses. It includes Zero Data Retention with Private Safety Processing, which runs automated safety reviews without OpenAI staff seeing the content. A Private Inference preview using confidential computing is due this fall.
The event landed under heavy safety scrutiny. CNBC reported that OpenAI had recently disclosed incidents of models behaving in unintended ways and had held back its latest Astra model (CNBC). The approval rules, read-only background research and monitoring built into dots should be read in that context.
What OpenAI DevDay 2026 means for web developers
- Recalculate your model costs. If you run GPT-6 Sol or Astra in production, GPT-6.1 Sol may give you similar quality at a much lower bill, especially with cached input.
- Treat ChatGPT as a distribution channel. Plugin extensions put your product in front of 1.2 billion weekly ChatGPT users.
- Prototype agents on hosted infrastructure. The Agents API with computer use removes much of the orchestration work you would otherwise build yourself.
- Use Decisions API for cheap routing. Constrained, fast answers are ideal for moderation, triage and agent control flow.
- Plan for agents as users. Dots and computer-use agents will visit your sites and apps, so clean markup, accessible forms and stable UI matter more than ever.
FAQ
When was OpenAI DevDay 2026? OpenAI DevDay 2026 was held on September 29, 2026 in San Francisco, with the keynote livestreamed on YouTube.
What was the biggest announcement at OpenAI DevDay 2026? Dots, always-on AI agents powered by GPT-6 Astra that work in the background with their own cloud computer.
How much does GPT-6.1 Sol cost? $2 per million input tokens, $0.10 per million cached input tokens and $10 per million output tokens.
Who can use OpenAI dots? ChatGPT Pro and Business Premium users in eligible markets, with an admin-enabled beta for Enterprise, Edu and Healthcare.
What is OpenAI Ultrafast? A premium speed tier offering up to 8x faster generation in Codex and up to 6x in the API, currently for GPT-6 Astra on Pro 500 and Enterprise.
Where can I watch the OpenAI DevDay 2026 keynote? The full keynote replay is on OpenAI's YouTube channel.
Sources
- DevDay 2026 Recap, OpenAI
- Introducing dots, OpenAI
- Introducing GPT-6.1 Sol, OpenAI
- Everything OpenAI Announced at DevDay 2026, BGR
- OpenAI DevDay 2026 live updates, CNBC
- OpenAI DevDay 2026 live blog, Engadget
- Live from OpenAI DevDay 2026: Keynote, YouTube
Also Read: Claude Opus 5.5 Released: Price Cut, Benchmarks & Breaking Changes for Developers
