ChatGPT Dots vs Claude Cowork

Choosing between ChatGPT dots vs Claude Cowork comes down to whether you need a dedicated agent monitoring business tools around the clock or an assistant producing structured deliverables inside conversational threads. Both systems keep running after you close your computer, yet they approach delegated work through distinct architectures.

Anthropic merged Cowork directly into Claude on September 16, 2026, rolling autonomous execution, document authoring, and slide generation into regular chat subscriptions. OpenAI introduced dots on September 29, 2026, giving subscribers an always-on cloud agent running on GPT-6 Astra with a dedicated browser and integrations across thousands of external services.

Dots launched today and are still rolling out, so real-world reports are not in yet, and everything said about dots comes from OpenAI's own announcement and help pages.

ChatGPT dots vs. Claude Cowork: Side-by-Side

Are you one of these vendors? Update your listing

The Difference in One Table

A ChatGPT dot is the better pick for standing work that watches your connected apps and reaches you in Slack or Teams; Claude, which absorbed Cowork on September 16, 2026, is the better pick when the work ends in documents and slides you edit together. Both keep working after you close your laptop.

Anthropic merged Cowork directly into Claude on September 16, 2026. The separate Cowork interface was integrated into the primary Claude product, meaning subscribers can trigger autonomous tasks, schedule recurring runs, and generate presentations directly within regular conversation threads without opening a separate utility.

OpenAI introduced dots on September 29, 2026, as distinct, standing AI coworkers powered by GPT-6 Astra. Each dot possesses its own cloud computer, its own cloud browser, and persistent memory that spans multiple sessions and integrated workplace applications. While Claude operates when assigned a specific objective, your dot acts as an enduring presence that conducts background research and checks connected software even while you are away.

• ChatGPT dots run as persistent identities on dedicated cloud machines with customizable names and avatars.
• Claude, which absorbed Cowork, works inside ordinary conversations and focuses on files and document outputs.
• Dots integrate directly into communication hubs like Slack and Microsoft Teams.
• Claude focuses deliverables into native authoring spaces through Claude Docs and Claude Slides.

Task Execution and Agent Architectures

Task execution separates into a persistent background agent model in ChatGPT dots and a goal-driven conversational delegation model in Claude.

OpenAI built dots to take on ongoing responsibilities across multiple projects at once. Each dot uses its cloud computer to browse the web, create tasks in Codex environments, and track updates in external apps. Through proactive research, a dot reviews connected platforms when you are not actively chatting, though OpenAI restricts these background routines to read-only tools that cannot alter data, send messages, or manipulate your browser. If you install the desktop app, you can optionally grant local computer access, allowing your dot to use local skills or your desktop browser when cloud access is blocked.

Claude works from a goal. You hand it a task, and it works across your files and tools, with a browser built into the side panel, and Anthropic says it keeps going after you close your laptop. You can also schedule it, as in Anthropic's example of a report Claude starts every Monday without being asked.

Deliverable formats highlight another major structural difference. Claude incorporates Claude Docs and Claude Slides directly into the workspace, allowing users to collaborate on presentations that export to PowerPoint or PDF. ChatGPT dots instead focus on communication and process coordination, interacting via chat, voice calls, Slack, or Microsoft Teams, while delegating complex document production to connected apps or ChatGPT Work.

Subscription Pricing and Usage Allowances

Claude offers a substantially lower entry price for autonomous workflows, while ChatGPT dots require premium subscriptions with higher base tiers. Software subscription prices adjust over time, so verify current figures directly on the ChatGPT pricing page and the Claude pricing page.

OpenAI does not list a dot in the Free, Go or Plus plans. Individual access requires a Pro subscription, which OpenAI structures as three monthly tiers: Pro 100 at $100 per month, Pro 200 at $200 per month, and Pro 500 at $500 per month. Annual billing is not available on Pro plans. For business teams, access requires a Premium seat in ChatGPT Business, priced at $125 per user monthly or $100 per user monthly on annual billing, with a two-seat minimum. OpenAI's rollout names Premium seats only; Standard Business seats cost $25 monthly or $20 annual.

OpenAI includes your first dot at no added subscription charge on Pro and Business Premium plans. OpenAI announced extended limits for the first month after launch, and says dots usage won't count toward eligible users' plan allowances for the next month. However, VentureBeat reported that OpenAI has not disclosed the numeric allowance for deeper work, the price for adding extra dots, or the cost of scaling a dot's monthly compute speed.

Anthropic makes Cowork features accessible on standard paid plans. Claude Pro costs $20 per month monthly or $17 per month billed annually ($200 up front) and includes Claude Code alongside Cowork capabilities. Higher-volume users can select Claude Max starting at $100 per month for 5x Pro usage. For organizations, Claude Team costs $25 per seat monthly ($20 annual) for Standard or $125 per seat monthly ($100 annual) for Premium, which includes five times standard usage and centralized administrative controls.

• Claude Pro provides autonomous background execution starting at $20 per month.
• ChatGPT dots require a minimum investment of $100 per month on Pro 100 or a Business Premium seat.
• Dots usage does not count against ChatGPT plan limits during the first month following launch.
• Anthropic provides 5x usage scaling on Claude Max ($100 per month) and Claude Team Premium ($125 per seat monthly).

→ Compare per-token pricing for every AI model

The Role of ChatGPT Work

ChatGPT Work represents the closest OpenAI counterpart to the document and slide authoring capabilities built into Claude.

OpenAI updated its agent strategy by retiring the former ChatGPT Agent and directing users to ChatGPT Work, which launched on July 9, 2026, The Next Web reported. While dots act as persistent, multi-channel coworkers, ChatGPT Work is an agent built specifically for longer, multi-step tasks that produce finished deliverables such as spreadsheets, reports, presentations, and Sites.

ChatGPT Work operates either in the cloud using a separate cloud browser or locally through desktop software with user consent. Dots do not replace ChatGPT Work. Instead, you can instruct your dot to initiate and supervise tasks inside ChatGPT Work or Codex, with those tasks consuming your standard workspace usage allowances.

A detailed breakdown of how OpenAI separates these two products is available in our comparison of ChatGPT dots vs ChatGPT Work. The older dot icons in the ChatGPT interface are unrelated to these agents; see what the ChatGPT dot means.

Security Controls and Supervisory Oversight

Both products ask before some actions, but OpenAI lets you write the rules, while Claude's default is to ask before taking an action.

OpenAI manages dot behavior through Custom Rules and an automated review system. Within the ChatGPT app, you can establish rule behaviors including taking action without asking, taking action if pre-approved in your prompt, asking before acting, or handing the task off to you entirely. Auto-review screens sensitive steps against your instructions, though OpenAI mandates that certain actions, such as changing passwords, always remain with human operators. Activity View shows your dot's progress, including background work, and lets you redirect it.

A dot creates memories from conversational feedback and connected apps. Disconnecting an application does not delete information the dot previously gathered; removing that data or clearing saved history requires resetting the dot, which deletes the dot, its conversations, saved memories and scheduled tasks. You can also pause your dot at any time through its profile menu.

Claude establishes oversight by asking for user approval before taking actions by default. Users who want unattended execution can adjust preferences to let Claude continue working independently, checking in only when a decision requires explicit review. On training, OpenAI says it doesn't use content from ChatGPT Business, Enterprise or Edu workspaces to improve its models by default; Anthropic's merge post and Cowork page do not address training.

• ChatGPT Custom Rules support four approval states: immediate action, pre-approved action, mandatory confirmation, or full handoff.
• Sensitive tasks like password modifications can never be delegated to dots.
• Resetting a dot is required to completely erase data obtained from previously connected applications.
• Claude defaults to asking before executing actions, with optional unattended permissions.

Regional Availability and Geographic Access

OpenAI publishes a region rule for dots; Anthropic's merge announcement describes the rollout by plan, not by region.

OpenAI restricts individual dots access based on geography. Dots are rolling out to Pro users in markets excluding the EEA, Switzerland and the UK. OpenAI has not published the underlying reasons for this territorial limitation.

Organizations have an alternative path. OpenAI says Business Premium users can get dots across all supported ChatGPT regions. Enterprise workspaces, including Edu and Healthcare, can try the beta once their workspace admin turns it on.

The merged Claude is rolling out across web, desktop and mobile to Pro and Max users, with Team and Free plans to follow soon. Enterprise admins will hear at least 30 days before any changes.

Operational Tradeoffs for Small Business Teams

Choosing an automation architecture depends on whether an organization needs continuous ambient monitoring or episodic project production. At Layer3Labs, we build and run automated workflows inside operating businesses, and we evaluate agent platforms based on where human oversight must sit.

In the client-intake and onboarding work we have automated for law firms, the job splits in two: watching for new inquiries and stalled steps, and producing the engagement letter. OpenAI's published design for dots, with read-only proactive research across connected apps and messages in Slack or Teams, lines up with the watching half; we have not run a dot on it, and real-world reports are not in yet.

Conversely, when the objective is producing client-facing engagement letters, synthesized case summaries, or presentation slides, Claude offers a more direct path. Built-in tools like Claude Docs and Claude Slides generate structured files that professionals can edit collaboratively inside the application. Both are in beta on paid plans, and Slides download as PowerPoint or PDF.

Organizations processing thousands of repetitive database records daily should not use either product. High-volume, programmatic data transformations belong in dedicated backend API pipelines or deterministic integration platforms, where per-task billing and execution limits are strictly controlled rather than tied to conversational seat licenses.

ChatGPT Dots vs Claude Cowork for Team Workflows

A workflow centered on communication and monitoring demands a different tool than a workflow centered on research and content generation.

Deploy ChatGPT dots if your team operates heavily inside Slack or Microsoft Teams and needs a persistent presence to track projects across multiple business systems. The ability to give your dot a persistent identity, connect it to over 4,000 apps, and rely on it to conduct proactive research makes it well suited for operational coordination. Keep in mind the minimum entry price of $100 per month and regional exclusions on personal accounts.

Pick Claude, which now includes Cowork, if your work ends in research reports, documents or slide decks. Claude Pro costs $20 per month, and Anthropic says Claude keeps working after you close your laptop, with Docs and Slides in beta on paid plans.

If most of your delegated work is watching for changes, try a dot on Pro or a Business Premium seat; if most of it ends in a document or deck, try Claude Pro first.

• Choose ChatGPT dots for 24/7 background app monitoring, Slack and Teams integration, and persistent conversational memory.
• Choose Claude, which now includes Cowork, for documents and slide decks, starting at $20 per month on Pro.
• Evaluate ChatGPT Work if your team requires OpenAI-based document generation rather than an ambient coworker.
• Audit team software permissions and administrative oversight settings before granting either agent access to external accounts.

The Verdict

ChatGPT dots and Claude Cowork address distinct operational needs. Choose ChatGPT dots if you require an always-on agent with its own cloud computer that monitors business software, conducts proactive background research, and interacts directly within Slack or Microsoft Teams. It represents a persistent operational presence, provided you are on Pro or a Business Premium seat and, for Pro, outside the EEA, Switzerland and the UK.

Choose Claude, which absorbed Cowork on September 16, 2026, if your delegated work ends in documents and slides. Claude Pro costs $20 per month, tasks can be scheduled, and Claude Docs and Slides are in beta on paid plans.

If your organization primarily handles structured, high-volume data transfers or backend synchronization, neither conversational agent is the appropriate choice. Teams in that situation should remain with dedicated workflow automation pipelines or direct API integrations that provide deterministic guarantees without conversational seat pricing. Running both also works: a dot can watch for changes while Claude drafts the documents. The answer shifts if OpenAI publishes the usage terms it has promised after the launch period, or if Anthropic brings the merged Claude to Team and Free plans.