LIVE from OpenAI DevDay: Dots, ChatGPT Space and GPT-6.1 Sol

The headline: ChatGPT becomes a place where agents work

OpenAI’s 2026 DevDay was not just another model launch. The company framed the event as its biggest DevDay yet, with more than 20 announcements across ChatGPT, Codex, models and new forms of AI-assisted work [1]. The center of gravity was clear: ChatGPT is being rebuilt as a shared surface where people, developers and agents collaborate, with OpenAI saying that surface now reaches 1.2 billion weekly users [1].

That number matters because it turns every DevDay announcement into a distribution story. A better model is one thing; a better model embedded inside the place where people already work is something else. OpenAI’s message to developers was that they are no longer just calling an API from the outside. They are being invited to build inside ChatGPT’s workflow layer, next to documents, coding tasks, plugins, meetings and agents [1].

The most visible new character in that system is the dot.

Dots: a personal agent, not just a smarter chat

OpenAI describes dots as “always-on” agents that learn what matters to a user, work on their behalf and take ongoing responsibilities off their plate [2]. Unlike a standard chat session, a dot has its own cloud computer and browser, can connect to more than 4,000 apps through OpenAI’s plugin ecosystem, and can be reached through ChatGPT, Slack or Microsoft Teams [2].

The product pitch is deliberately personal. A dot can monitor feedback, prepare pull requests, update launch materials, rerun analysis, revise proposals or process content production workflows, while the human stays in the approval loop [2]. OpenAI says users begin with a primary dot, with teams of dots envisioned later [2].

The rollout is also narrower than the 1.2 billion-user headline. Dots are beginning across Pro, Business Premium and Enterprise plans in eligible markets, while Enterprise, Edu and Healthcare workspaces can test the beta only when an admin enables it [2]. OpenAI says the first dot is included with Pro or Business Premium, conversations with the dot do not count toward ChatGPT usage limits, and deeper Codex or ChatGPT Work tasks still count as usual [2].

This is the key distinction: dots are meant to feel conversational, but they are really an interface to persistent work. The risk is equally obvious. A bot that can keep researching, using apps and preparing actions needs guardrails that are stronger than the old “the model might hallucinate” warning.

ChatGPT Space: the agent office

Dots are not landing alone. ChatGPT Space is the workspace around them. OpenAI describes Space as a dedicated home where teammates, ChatGPT and a user’s dot can build on shared knowledge, with ChatGPT helping keep the space organized from user instructions [1]. Pages, another new surface, are collaborative documents where people and agents can write, research, generate charts, create images and visualize information together [1].

That is a meaningful product shift. ChatGPT is no longer only a question box; it is becoming a project container. Add collaborative slides, shared team tasks, Slack and Teams mentions, meetings notes, plugins, Sites and shareable profiles, and the strategic direction is obvious: OpenAI wants the workplace to happen around ChatGPT, not merely consult ChatGPT [1].

This also explains why DevDay felt less like a single breakthrough and more like a platform bundle. Codex in the cloud lets developers run tasks from a computer, phone or cloud environment; the Agents API gains computer use; plugin extensions let developers build interactive panels and file viewers inside ChatGPT; MCP events let connected apps trigger automations [1]. Each piece makes the platform more “agent-ready.”

GPT-6.1 Sol: cheaper frontier work, not the flashiest model

The model announcement was GPT-6.1 Sol, an upgrade to GPT-6 Sol that OpenAI says nearly matches GPT-6 Astra’s intelligence for agentic coding, computer use and professional work at one-fifth of Astra’s standard input and output token prices [3]. In practical terms, this is the model designed to make persistent agents economically plausible. If agents keep context, retry tasks and run workflows over hours, cost per unit of useful work becomes just as important as leaderboard prestige.

OpenAI says GPT-6.1 Sol is available in ChatGPT Work and Codex for Plus, Pro, Business, Enterprise and Edu users, and through the API as gpt-6.1-sol; it is not yet available in regular Chat [3]. Standard API pricing is listed at 2 dollars per million input tokens, 10 cents per million cached input tokens and 10 dollars per million output tokens [3]. OpenAI is also preparing an Ultrafast option for GPT-6.1 Sol, while GPT-6 Astra Ultrafast is already part of the DevDay package for selected plans and API use [1].

Safety was part of the model story, not an appendix. OpenAI’s system-card addendum says GPT-6.1 Sol is treated as Critical in cybersecurity and High for biological and chemical capability under its Preparedness Framework, with the same safeguards stack as GPT-6 Astra [5]. The same document says GPT-6.1 Sol reaches the Critical threshold for cybersecurity capabilities under OpenAI’s internal assessment [5].

That is a double-edged message. OpenAI is saying the model is useful enough to power serious agents, and risky enough to require frontier-model controls.

The safety cloud over DevDay

The DevDay timing made that tension impossible to ignore. AP reported that Sam Altman introduced the new always-on agent a day after OpenAI halted the rollout of a more advanced model over safety concerns [4]. Altman did not dwell on that shelved model in the keynote, but said during a Q&A that OpenAI was investing more in safety, security and monitoring of AI agents [4].

That context changes how dots should be read. The product is not just a cute assistant. It is a bet that users will allow AI systems to remain active in the background, inspect information, use tools and prepare consequential work. OpenAI’s dots post emphasizes boundaries, app permissions, activity views, action approvals, read-only proactive research and monitoring that can pause or stop a dot when safety concerns arise [2].

The harder question is whether users will experience those controls as clarity or friction. If dots ask permission too often, they become glorified reminders. If they ask too rarely, they become a governance headache. DevDay’s real test begins after the livestream, inside messy workspaces with private data, conflicting instructions and impatient humans.

The market around the keynote

The broader AI market gave DevDay an extra charge. AMD announced on September 28 that it had agreed to acquire World Labs, the AI model and research lab led by Fei-Fei Li, in an all-stock transaction valued at about 8.2 billion dollars and expected to close by the end of 2026, subject to approvals [6]. The deal reinforces the same theme OpenAI pushed from the stage: the next phase of AI will be shaped by the coupling of models, software, hardware and workflows.

Meanwhile, Oura postponed its IPO despite momentum from the Oura Ring 5, 5.7 million paid members and expected fiscal-year revenue growth of 90 percent [7]. That delay shows how uneven the AI-adjacent market remains: investor excitement is real, but timing, valuation and macro confidence still matter.

Axios captured the contradiction around the week: AI leaders were showing off powerful new systems while warnings about the technology’s dangers continued to swirl [8]. That is the background music for OpenAI’s platform push. The industry wants agents to unlock productivity, but it has not yet proved that always-on AI will create more signal than noise.

Why this DevDay matters

The practical takeaway is simple: OpenAI is trying to make ChatGPT the operating layer for agentic work. Dots are the user-facing metaphor. ChatGPT Space is the shared room. GPT-6.1 Sol is the cost-performance engine. Codex, plugins, Pages, team tasks and the Agents API are the developer surfaces.

If it works, the next platform war is not only about who has the smartest model. It is about who owns the environment where agents receive context, coordinate with humans, use tools and return completed work. DevDay 2026 was OpenAI’s clearest answer so far: that environment should be ChatGPT.

The skeptical reading is just as important. More agents can mean more output, more notifications, more half-finished automations and more low-value AI content dressed up as productivity. The difference between useful delegation and industrial-scale slop will come down to permissions, review design, pricing, reliability and whether users can tell what their agents are doing.

OpenAI has moved from “ask me anything” to “let me handle it.” DevDay showed the ambition. The next few months will show whether dots become trusted coworkers, expensive toys or the first normal form of always-on AI work.

Sources from the last 72 hours

1. [1]DevDay 2026 RecapSep 29, 2026, 2:00 AM
2. [2]Introducing GPT-6.1 SolSep 29, 2026, 2:00 AM
3. [3]Addendum to GPT-6 Astra System Card: GPT-6.1 SolSep 29, 2026, 2:00 AM
4. [4]AMD to Acquire World Labs to Advance the Future of AI ComputeSep 28, 2026, 2:00 AM
5. [5]The AI industry's contradictions take center stage in Washington, Silicon ValleySep 29, 2026, 11:00 AM
6. [6]Introducing dotsSep 29, 2026, 2:00 AM
7. [7]OpenAI CEO Sam Altman announces agent Dots, GPT-6.1 Sol and other updatesSep 29, 2026, 8:51 PM
8. [8]Investors will have to wait for Oura's IPOSep 29, 2026, 3:57 PM

AI-generated article based on recent web research, then preserved as a dated editorial snapshot.