🔍 Read the full analysis: Key Questions Raised By OpenAI Dev Day 2026 on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
At DevDay on 29 September 2026, OpenAI announced more than 20 products and features positioning ChatGPT as a platform shared by 1.2 billion weekly human users and AI agents. Dots, plugin upgrades, the Decisions API and Sign in with ChatGPT raise pressing questions about agent reliability, plugin economics and platform dependence.
OpenAI used its DevDay on 29 September 2026 to shift its positioning from chatbot maker to agent platform operator, announcing more than 20 products and features and pitching ChatGPT as a shared surface for humans and agents reaching 1.2 billion weekly users, according to the company. The headline items — always-on agents called Dots, a reworked plugin ecosystem, a limited-preview Decisions API and Sign in with ChatGPT — confirm a strategy in which agents, not conversations, become the primary unit of work. But several of the announcements depend on previews, pilots and promises, leaving builders and buyers with questions about reliability, pricing and lock-in.
Dots are OpenAI’s biggest consumer bet: persistent agents powered by GPT-6 Astra, each running on its own cloud computer and browser, able to reach over 4,000 apps through plugins. According to OpenAI, a dot works toward a user’s goals around the clock, learns preferences over time and can keep several projects moving at once. Access is open now to Pro and Business Premium users in eligible markets, with beta availability for Enterprise, Edu and Healthcare once an admin enables it. The first dot is included in the plan, and chatting with it does not count toward usage limits, though tasks it starts in Codex or ChatGPT Work do.
The plugin ecosystem received what OpenAI described as a major upgrade: plugin extensions that live in the ChatGPT sidebar, a new Plugin Creator and submission flow, improved discovery, Sites that host plugins for teams, and support for the proposed MCP Events specification so plugins can trigger automations when something happens in a connected app. Figma and Adobe showed integrations, and an OpenAI Marketplace lets eligible enterprise customers spend part of their OpenAI commitment with 32 launch partners, including Salesforce, ServiceNow, Harvey and CrowdStrike.
Two further announcements round out the platform pitch. The Decisions API, in limited preview, applies GPT-6 Luna to narrow, repetitive choices: developers define questions with a finite list of allowed answers, supply text or images as context, and receive a selection usable for classification, routing or choosing an agent’s next action. The Agents API now supports computer use, letting agents operate software through its interface, with OpenAI running the Codex harness on the developer’s behalf. On the model side, OpenAI announced GPT-6.1 Sol, which it described as near-Astra intelligence at one-fifth of Astra’s token prices, alongside an Ultrafast tier and Pro 500.
OpenAI DevDay 2026 at a glance
The five launches that matter, plus Sol
The rest of the slate
Collaboration
ChatGPT SpacelivePagesliveCollaborative slidesnext few weeksTeams and Team Taskslive@ChatGPT in Slack and TeamsliveMeetings pluginbeta, macOSShareable profilesliveCodex
Codex CloudliveCLI refreshvoice, /agentsCode ReviewliveCodex Security CloudlivePlans, enterprise, speed
Pro 500$500, 25× PlusUltrafastAstra now, Sol soonPrivate IntelligencePrivate Inference this fallBedrock Managed Agentsruns in AWSOpenAI Marketplace32 partners, applyWhy Plugin Selection Now Decides Who Gets Hired
The commentary around the event, including analysis published by ThorstenMeyer, makes the strategic stakes explicit: an agent picks a business by outcome, not by brand. A customer asks for a result, ChatGPT understands the job, and a plugin is selected to do the work — meaning, in the analysis’s framing, a business can be hired before the customer knows its name. That shifts competition toward measurable plugin performance: whether the agent’s intent matched, whether activation occurred, and whether the job completed.
For software builders, distribution economics change. If dots reach thousands of apps only through plugins, then plugin visibility becomes the new app-store ranking. Developers who ignore the directory risk being invisible to agent-driven demand; those who invest face unclear revenue terms, evolving review processes and the risk that OpenAI’s own discovery layer, not user choice, determines who wins. Sign in with ChatGPT, live with 16 launch partners, makes a free-core, paid-upsell model workable by letting users bring existing plans — which deepens dependence on OpenAI as both identity provider and distribution channel.
Popular right now
Ninja AF101 4-Quart Air Fryer, 4-in-1, Grey
Hisense 40-Inch QD4SR Series Hi-QLED FHD Roku Smart TV
eufy E25 Robot Vacuum and Mop with HydroJet Roller Mop
As an affiliate, we earn on qualifying purchases.
From Chatbot to Agent Platform in One Event
DevDay 2026 continues OpenAI’s multi-year expansion from a consumer chat product into developer infrastructure. Previous DevDays introduced function calling, GPTs and early agent tooling; this event consolidates those threads into a single agent stack: models (Astra, Sol, Luna), runtimes (dots’ cloud computers, the Codex harness), distribution (the plugin directory and Marketplace), identity (Sign in with ChatGPT) and governance (Microsoft Agent 365 integration planned for specialist dots).
Control mechanisms announced alongside Dots signal OpenAI’s awareness of agent risk. Each dot runs on an isolated cloud computer; saved passwords can be used without being exposed to the model; background research is limited to read-only tools; Custom Rules let users allow, require approval for, or block specific actions; and an Activity View plus monitoring can pause or stop a dot. Sensitive tasks such as password changes stay with the user. Specialist dots — enterprise agents with their own identity, credentials and IT-provisioned hardware — begin with pilots in procurement, invoice processing, email marketing, customer support and commercial contracting. Notably, the developer docs include a guide for submitting a Claude Code plugin, suggesting OpenAI intends to accept skills built for competing agent ecosystems.
What OpenAI Did Not Yet Answer
Several announcements rest on previews or unverified claims. The Decisions API is in limited preview, with broad release promised but no firm date. Dots’ deeper capabilities — texting support, teams of dots working together, add-on dots with extra speed and output — are described by OpenAI as future features. One widely circulated dot anecdote, in which an early tester’s dot noticed an unsubmitted invoice, prepared it and sent it after approval, is a company-provided example rather than an independently verified case.
Reliability remains the central open question. OpenAI itself warns that dots can make mistakes, and it is not yet clear how the approval, Custom Rules and monitoring controls perform under real workloads. Pricing beyond the included first dot is only sketched. The near-Astra claim for Sol and the value of the Marketplace’s 32 partners have not been tested by third parties. And the strategic bet that consumers will delegate always-on agents to act on their behalf — including sending invoices — is unproven at scale.
Rollouts, Pilots and the Marketplace Race
Watch for the Decisions API’s broad release and its published pricing; expansion of dot access beyond Pro and Business Premium; texting support and multi-dot teams; and the first results from specialist dot pilots in procurement, support and contracting, including the planned Microsoft Agent 365 governance integration. For builders, the near-term milestones are practical: getting plugins through the new submission flow, tracking how directory ranking and in-conversation recommendations affect activation, and deciding whether to support MCP Events automations. The bigger question — whether agent-mediated selection becomes a dominant channel for software demand, and on whose terms — will play out over the coming quarters as usage data from the 1.2-billion-user surface accumulates.
Key Questions
What are Dots in OpenAI’s DevDay 2026 announcements?
Dots are always-on personal agents powered by GPT-6 Astra, each with its own cloud computer and browser, connecting to over 4,000 apps via plugins. They are available now to Pro and Business Premium users in eligible markets, with beta access for Enterprise, Edu and Healthcare organizations.
Is the Decisions API generally available?
No. It is in limited preview, with a broad release promised but not yet dated. It uses GPT-6 Luna to pick from a finite list of answers you define, for tasks like classification and routing.
How much do Dots cost?
According to OpenAI, the first dot is included in the plan, and chatting with it does not count toward usage limits. Tasks a dot starts in Codex or ChatGPT Work do count. Paid add-ons — extra dots, more speed and monthly output — are planned for later.
Why do plugins matter after DevDay 2026?
Plugins are now the way into ChatGPT, Codex and dots. With over 4,000 apps reachable through the ecosystem, plugin distribution largely decides which services receive agent-driven work, making directory ranking and conversion performance a new competitive battleground.
What safeguards exist for always-on agents?
Each dot runs on an isolated cloud computer, background research is read-only, Custom Rules let users gate specific actions, and monitoring can pause or stop a dot. Sensitive tasks like password changes stay with the user. OpenAI still cautions that dots can make mistakes.
Source: ThorstenMeyerAI.com
