Construct vs ChatGPT Dots

Construct Team

Construct Team · Editorial team

The team building and documenting Construct Computer.

Published · @use_construct

ChatGPT Dots are OpenAI's always-on agents, and your first dot comes with a ChatGPT Pro subscription ($100 or $200 a month, depending on which launch report you read) or Business Premium; there is no Free or Plus access. Construct is an AI employee with its own persistent workspace, starting at $9/month on Lite, with bring-your-own-key model access on Pro.

Choose Dots if you already pay for ChatGPT Pro, want the largest app catalog, and value approvals and custom rules built in from day one. Choose Construct if you want a lower entry price, a choice of model providers, memory you can inspect and correct, and a workspace that holds files, schedules, and private internal apps.

Try Construct

What is ChatGPT Dots?

OpenAI launched dots at DevDay on September 29, 2026. A dot is an agent inside ChatGPT that you give a goal; it works toward that goal on its own cloud computer and browser, using the apps you connect, and brings the result back for review. OpenAI calls them "remarkably capable, always-on agents built to handle everything," according to Decrypt's DevDay roundup.

Launch coverage describes these capabilities:

  • Its own computer. Each dot runs on GPT-6 Astra with a dedicated cloud computer and browser, and can work on several projects at once (The Next Web).
  • A large app catalog. Dots reach more than 4,000 apps through OpenAI's plugins (SiliconANGLE).
  • Many places to talk to it. You can message or call a dot in ChatGPT on desktop, web, and mobile, or message it in Slack and Microsoft Teams, with text messaging listed as coming soon (The Next Web).
  • Background work. When you are not working with it, a dot does "proactive research" with read-only app access, so it cannot send messages or change accounts during that time (SiliconANGLE).
  • Specialist dots. Employers can provision single-purpose dots with their own identities and credentials, and Microsoft is integrating them with Agent 365 controls (TechCrunch).

Dots are a different product from chatting with ChatGPT. If your question is how Construct compares with the ChatGPT chat assistant itself, read Construct vs ChatGPT, Claude, and Gemini. This page is about the agent.

Methodology

We build Construct. This is our editorial comparison of documented capabilities and published prices, checked on October 1, 2026. It is not a hands-on performance benchmark.

OpenAI's own announcement and help pages refused automated requests when we checked, so the Dots facts below come from launch coverage by SiliconANGLE, The Next Web, Decrypt, Fortune, TechCrunch, and BGR, published September 29 and 30, 2026. Where those reports disagree, we show both versions. Where a capability is not documented in the sources we reviewed, we say so rather than treating missing documentation as proof that it is absent.

Construct vs ChatGPT Dots: features at a glance

Construct vs ChatGPT Dots capabilities
CapabilityConstructChatGPT Dots
Work environmentPersistent workspace with files, live browser runs, sandbox terminalDedicated cloud computer and browser per dot; laptop access with permission
ModelBundled usage; BYOK for OpenRouter, OpenAI, Anthropic, Bedrock, xAI on ProGPT-6 Astra; no alternative provider documented in reviewed sources
AppsLive integration catalog plus custom MCP toolsMore than 4,000 apps through OpenAI plugins
Safety controlsSupervised runs; linear workflows without approval gates todayApproval checkpoints, Custom Rules, built-in rules, and a monitoring system
MemoryInspect, correct, or forget stored contextLearns preferences from feedback; inspect and edit controls not documented
Recurring workSaved linear workflows and Calendar schedules (3 on Lite)Always-on work toward goals, plus read-only proactive research
Internal appsBuild and open private workspace appsChatGPT Space for shared documents, spreadsheets, and slides
Ways to reach itWeb, Slack, Telegram, Discord slash commands, native agent inboxChatGPT desktop, web, and mobile (message or call), Slack, Teams; SMS coming soon

Dots sources: SiliconANGLE, The Next Web, and Fortune. Construct's product guide explains its execution surfaces and limits.

How much does ChatGPT Dots cost?

There is no standalone Dots price. Your first dot is included with an eligible ChatGPT subscription, and OpenAI has promised paid options for more dots, more speed, or more capacity later at prices it has not announced (SiliconANGLE). These are USD subscription prices billed monthly, without annual discounts.

Monthly subscription prices for Construct and ChatGPT Dots
SubscriptionMonthly priceWhat to budget for
Construct Lite$9Up to 2 agents, 50 steps per task, 3 scheduled tasks, agent email
Construct Starter$596x Lite usage, up to 5 agents, 150 steps per task, 10 scheduled tasks
Construct Pro$29932x Lite usage, up to 15 agents, up to 1,000 steps per task, BYOK
ChatGPT Pro 100$100Reports disagree on whether a dot is included (see below)
ChatGPT Pro 200$200First dot included
ChatGPT Pro 500$500Highest included usage and Astra Ultrafast; first dot reported included
ChatGPT Business PremiumPer seatFirst dot included

Sources: Construct pricing, Pulse 2.0 on the three Pro tiers, BGR, and Carly's Dots guide. Subscriptions and limits can change; check both before purchasing.

The sources disagree on the entry price. BGR reports that dots "are only available on ChatGPT's Pro 200 and Business Premium plans" and are "not currently available on the cheapest of the Pro tier, which costs $100 a month." Carly's guide lists Pro 100 as including a dot. Fortune says all Pro and Business Premium users get one dot but only discusses the $200 and $500 plans. So the cheapest personal route to Dots is either $100 or $200 a month. Until OpenAI's own page settles it, budget for $200.

Usage is also still open. During the first month, dots usage does not count against plan allowances, and OpenAI has not published the terms that apply after that (SiliconANGLE, Traictory). Conversations with a dot do not count toward ChatGPT usage limits, but tasks it starts in Codex or ChatGPT Work do (The Next Web). Pro subscribers in the European Economic Area, Switzerland, and the UK are excluded for now.

What this means in practice:

  • Starting from zero, Construct Lite is $91 to $191 a month less than the cheapest reported route into Dots. That is a lower entry cost, not proof of equal usage or a lower cost per finished task.
  • If you already pay for ChatGPT Pro, your first dot costs nothing extra. Trying it before adding another subscription is reasonable.
  • At the top end, Construct Pro at $299 costs more than ChatGPT Pro 200. Compare it with Pro 200 only if you need what Pro adds: BYOK, up to 15 agents, and up to 1,000 steps per task.

Try Construct

Where does Construct give you more control?

Both products give an agent a computer. The differences show up in who supplies the model, what you can see and change about the agent's context, and where the finished work lives.

Pick the model, or bring your own keys

Launch coverage describes Dots as powered by GPT-6 Astra, and we found no documented option to run a dot on another provider's model. That is a sensible default if you like OpenAI's models.

Construct Pro supports your own keys for OpenRouter, OpenAI, Anthropic, Amazon Bedrock, and xAI. BYOK can be off, used as a fallback after bundled limits, or set for exclusive use, as the Construct terms explain. You can use OpenAI through your own key and switch to another provider later without moving your work. BYOK adds provider charges when your keys are used, and it requires Pro.

Correct what the agent remembers

Fortune reports that dots learn your preferences over time and remember past conversations and decisions. The sources we reviewed do not document a way to see, correct, or delete specific learned facts.

Construct lets you inspect stored memory, correct a fact, or forget it. If a client changes its reporting format, you update the context the next job relies on instead of hoping the agent notices. The memory guide covers how that works. Activity summaries and bounded tool records help you investigate a run; they are useful operating records, not a complete immutable audit log.

Keep the work in a workspace, not a thread

A Construct run saves finished work to the workspace as it is produced. If a run stops at step 30, steps 1 to 29 are still there, and the next run picks up from what exists. Files, schedules, and outputs sit in one place you can open.

When several people review repeated output, Construct can build a private workspace app, such as a review queue with owners and statuses. The internal-tools guide explains the build lifecycle and its limits. Dots has its own collaboration surface: Fortune describes ChatGPT Space, where people and agents edit documents, spreadsheets, and slides together. The difference is that Construct can build a small purpose-made app, not only share documents.

Reach it where your team already talks

Dots covers ChatGPT, Slack, and Teams. Construct covers the web, Slack, Telegram, Discord slash commands, and a native agent email address that can receive and send mail. If your community lives on Discord or your team coordinates on Telegram, Construct reaches you there. If your company runs on Teams, Dots does and Construct does not.

Stay independent of one model vendor

Dots lives inside ChatGPT, and access depends on your ChatGPT plan and region. Construct is a separate workspace that can use several model providers. For a small team that wants to keep its procedures, files, and memory independent of any one lab's roadmap, that separation is the point.

Running the same job on both

We have not run this test yet. When we do, we will publish the prompt, the time to a usable result, the cost, and what each product left behind.

Here is how the Construct side of that job would run. This is an illustrative walkthrough, not a measured result.

Save the three competitor URLs and the column layout in a file in the workspace. Ask the agent to visit each pricing page in a live browser run, write a dated CSV, and compare it with last week's file. Check the first file by hand. Once it is right, save the steps as a workflow and put it on the Calendar; Lite includes 3 scheduled tasks, which is enough for this one. Keep delivery out of the job until you trust the numbers. The workflow guide explains the current boundaries: workflows are linear, with no branching, delays, or approval gates.

Where is Dots the better choice?

Dots wins in several places, and they matter.

  • App catalog. More than 4,000 apps through plugins is a large catalog. If your stack is long and mainstream, Dots is more likely to connect to all of it without custom work.
  • Approvals by default. Dots checks actions that affect your accounts or share information, and Custom Rules let you allow, block, or require approval for specific actions (The Next Web). Jobs like changing a password always stay with the human (Decrypt). Construct's workflows have no approval gates today, so steps with irreversible effects, such as a payment or a customer email, need your supervision.
  • Proactive help. A dot can notice work you did not assign. Developer Dan McAteer told The Next Web: "My dot proactively picked up on an invoice I needed to send for my freelance writing, pulled the relevant details from the email thread and drafted the invoice. After I approved it, dot sent the invoice as a PDF."
  • Teams and voice. Microsoft Teams support and the option to call your dot are not part of Construct's documented channels.
  • Distribution. If your team already lives in ChatGPT, there is nothing new to adopt, no second login, and no second bill.

Moving from Dots to Construct

Start with one procedure, not the whole setup. Copy the goal and instructions you gave your dot, move the source files into the Construct workspace, and reconnect the apps the job needs. Run it once by hand and check the output before you schedule it.

Do not expect an import. There is no tool that moves a dot's learned preferences, rules, or connected accounts into Construct. Write down the rules you set in Dots, especially which actions needed approval, and keep those steps supervised in Construct, since its workflows do not pause for approval.

Which should you choose?

Choose Dots if you already pay for ChatGPT Pro or Business Premium, your team works in ChatGPT and Teams, and you want built-in approval rules plus the largest app catalog. It is a strong default for people who are happy with OpenAI's models.

Choose Construct if you do not want a $100 to $200 subscription to start, if you want to pick or bring your own model provider, or if your recurring work needs a place to live: files, schedules, correctable memory, and small internal apps. It also fits teams that work in Telegram, Discord, or email.

Some teams will run both: a dot for personal follow-ups inside ChatGPT, and Construct for the recurring team jobs whose files and history need to outlast any one conversation.

Give Construct one real job

Pick the report, price check, or tracker that keeps slipping into your evening. Give Construct the inputs and describe what a finished result should contain. Start with Lite at $9 for a small task, or use the seven-day Pro trial to test BYOK and longer runs. Check the current plans against your workload, then judge the saved result and the effort it took to review it.

Sources

Related resources

Frequently asked questions

How much does ChatGPT Dots cost?
There is no standalone Dots price. Your first dot is included with an eligible ChatGPT Pro or Business Premium subscription, and launch reports disagree on whether the $100 Pro 100 tier qualifies or whether you need Pro 200 at $200 a month. OpenAI has not announced prices for extra dots or the usage terms that apply after the first month.
Is there a free version of ChatGPT Dots?
No. Launch coverage says Dots are not available on Free, Go, or Plus plans, and Pro subscribers in the European Economic Area, Switzerland, and the UK are excluded for now.
What is a cheaper alternative to ChatGPT Dots?
Construct starts at $9 a month on Lite, with up to 2 agents, 50 steps per task, 3 scheduled tasks, and an agent email address. That is a lower entry cost than the $100 or $200 ChatGPT Pro route into Dots, not proof of equal usage or a lower cost per finished task.
Can I choose the AI model in ChatGPT Dots or Construct?
Launch coverage describes Dots as powered by GPT-6 Astra, and we found no documented option to use another provider. Construct Pro supports your own keys for OpenRouter, OpenAI, Anthropic, Amazon Bedrock, and xAI.
Where is ChatGPT Dots better than Construct?
Dots connects to more than 4,000 apps, checks sensitive actions for approval by default with Custom Rules you can set, works in Microsoft Teams, and is included if you already pay for ChatGPT Pro. Construct's workflows have no approval gates today, so irreversible steps need your supervision.
Can I import my dot into Construct?
No. There is no tool that moves a dot's learned preferences, rules, or connected accounts into Construct. Start with one procedure, move its instructions and files into the workspace, reconnect the needed apps, and test it by hand before scheduling it.

Try Construct

  1. Construct vs ChatGPT, Claude, and Gemini

    comparison · Jul 20, 2026 · Construct Team

    Compare Construct with chat assistants such as ChatGPT, Claude, and Gemini across workspace, tools, memory, schedules, and workflows.

  2. Chat Assistants vs AI Employees

    article · Jun 15, 2026 · Ankush

    ChatGPT and Claude are great at drafting, but autonomous ops need execution, persistence, and inspectable work. See how Construct compares to chat assistants.

  3. AI employees for small business: 10 platforms compared

    comparison · Sep 29, 2026 · Construct Team

    The best AI employee platforms for small business in 2026, compared on price, capabilities, and limits: Lindy, Sintra, Marblism, Manus, Construct, and more.

  4. Grok Bot Alternative: Construct vs Grok Bot

    comparison · Sep 10, 2026 · Construct Team

    Compare Construct and Grok Bot on monthly pricing, workflows, integrations, memory, and model choice for small teams.