Key Notes
- Dots introduces persistent personal agents, while GPT-6.1 Sol expands model choices for coding and professional work.
- Astra Ultrafast and the $500 Pro tier add speed and capacity, with separate usage costs and access rules.
- ChatGPT Space, Codex Cloud and plugin extensions broaden how people and agents work together.
OpenAI used DevDay 2026 to introduce Dots, upgrade its Sol model and put a much faster version of Astra behind a new premium subscription. The September 29 event also expanded ChatGPT’s role as a place to collaborate, run software and hand ongoing work to AI agents.
The company listed more than 20 major announcements in its event recap. The most consequential changes fall into three areas: agents that keep working between conversations, more choices over model cost and speed, and tools that bring people and applications into the same workspace.
Dots: Personal Agents That Keep Working
Dots are the clearest expression of that shift. Powered by GPT-6 Astra, each agent has a cloud computer and can use connected applications to carry out assignments over time. Users can name their dot, give it context and reach it through ChatGPT, Slack or Microsoft Teams.
OpenAI’s launch announcement describes agents that learn preferences from feedback and handle several projects at once. The proposed uses range from preparing content to investigating bugs and updating plans as new information arrives. These are capabilities OpenAI is presenting at launch, rather than evidence that every workflow can run without supervision.
The rollout starts with Pro and Business Premium in eligible markets. Enterprise, Edu and Healthcare users can try the beta when an administrator enables it. Conversations with a dot do not count toward ChatGPT usage limits, but tasks it starts or manages in Work and Codex use the normal allowance.
Users also retain controls over connected apps and actions. OpenAI says proactive background research uses read-only tools. AIstify’s Dots coverage explains the product and its access rules in more detail.
GPT-6.1 Sol: More Capability at Sol Prices
GPT-6.1 Sol arrives one week after GPT-6 Sol, with OpenAI positioning it closer to Astra on demanding coding, document and computer-use tasks. The pitch is stronger performance without moving customers to the flagship model’s standard token prices.
Standard API input and output cost $2 and $10 per million tokens respectively, compared with Astra’s $10 and $50. Those rates make Sol one-fifth the price on those two measures; they do not guarantee that every completed job costs exactly one-fifth as much. Reasoning, repeated attempts and tool use can change the total.
OpenAI’s model guidance places GPT-6.1 Sol in ChatGPT Work and Codex, rather than regular Chat. The rollout covers eligible Plus, Pro, Business, Enterprise and Edu accounts, alongside API access. AIstify’s Sol breakdown includes the pricing comparison and evaluation results.
Ultrafast: Faster Output With a Usage Trade-Off
Astra Ultrafast adds another way to pay for performance. OpenAI says it generates tokens up to eight times faster than Standard Astra in Codex, with its DevDay announcement citing 300 tokens per second. The API comparison is up to six times faster.
The distinction between generation and task completion matters. A coding agent may still need to inspect a repository, run tests or wait for a service. Faster output can reduce the pauses between those steps, but it does not make every end-to-end assignment eight times faster.
OpenAI’s speed documentation also sets out the cost: Astra Ultrafast uses included subscription allowance at eight times the Standard rate, while purchased credits are charged at six times the Standard credit rate. Access in Work and Codex starts on Pro 500 and eligible Enterprise and Edu plans. GPT-6.1 Sol Ultrafast is coming later.
Pro 500: A New Top Personal Subscription
The new $500 monthly ChatGPT Pro tier combines Ultrafast access with an allowance advertised at 25 times that of Plus. It joins the existing $100 and $200 Pro options, giving heavy users another subscription choice without making Ultrafast available across every paid plan.
The launch coincides with revised allowances for Pro 200. OpenAI’s plan guidance says eligible existing subscribers can keep their previous allowance through October 29 before moving to the lower level. Buying credits on Pro 100 or Pro 200 does not unlock Ultrafast at launch. AIstify’s Pro 500 guide covers the trade-offs.
ChatGPT Space: Shared Pages for People and Agents
ChatGPT Space brings files and collaborative pages into a shared working environment. Pages can combine writing, research, charts and interactive content, with colleagues editing together and asking ChatGPT or a dot to make changes.
Space is available on Pro, Business and Enterprise. Page creation and editing work on the web and desktop; mobile users can find, read and share pages, with mobile editing still to come. Collaborative slides and spreadsheets are also described as forthcoming features.
This could reduce the handoff between generating an answer and maintaining a working document. A team can keep revising the same plan or dashboard as its inputs change, rather than repeatedly copying fresh responses into another application. Its usefulness will depend on how well shared context and access controls fit the team’s existing workflow.
Codex Cloud: Coding Work That Outlasts a Laptop Session
The expanded Codex Cloud workflow lets developers start coding assignments in hosted environments and review or continue them from the web, mobile or desktop. A task can keep running when the user’s own computer is asleep.
Reusable environments hold the repositories, dependencies and tools a project needs, while individual tasks get their own workspace. Developers can inspect changes, request follow-ups and turn the results into commits or pull requests. The practical benefit is continuity: a long investigation or implementation does not have to occupy a local session from start to finish.
Plugins and Sign-In Extend ChatGPT Beyond the Chat Window
OpenAI also introduced plugin extensions that can add sidebar applications, panels beside conversations and dedicated file viewers. These give developers more room to build an interface around their service. Availability varies by feature and client; the documentation says web extensions for Free and Go are coming soon.
A separate sign-in expansion lets Plus and Pro subscribers authorize participating apps to use their ChatGPT plan for eligible AI requests. That usage draws from the shared Work and Codex allowance. Users can set a weekly limit for each app, and third-party subscriptions or service charges may still apply.
What the Announcements Add Up To
DevDay’s central theme was continuity. Dots can carry responsibilities between conversations, Space gives collaborators a shared place to work, and cloud coding environments let tasks continue away from a laptop. Sol and Ultrafast provide different choices over the cost and speed of that work.
These products also make the subscription question more complicated. Access, included usage, speed settings and rollout status are separate considerations. For customers, the useful test will be whether the new tools complete more worthwhile work with less coordination, at a cost they can predict.
Disclaimer: AIstify is an independent media brand owned and operated by NuvexMedia LLC, publishing news, research, and insights on artificial intelligence, emerging technologies, automation, and related industries. NuvexMedia LLC invests in and collaborates with companies across the AI, technology, software, and digital innovation sectors. These relationships do not influence AIstify’s editorial coverage, and the publication maintains full editorial independence to provide accurate, timely, and objective information. © 2026 NuvexMedia LLC. All rights reserved. This content is for informational purposes only and should not be considered legal, tax, investment, financial, or other professional advice.
