ChatGPT vs Claude Cowork: What Counts as Finished Work?
Compare ChatGPT and Claude Cowork on completion criteria, code review, recurring triggers, memory and the scope of local execution.
Founder, Task Machine
ChatGPT vs Claude Cowork becomes a completion question when your team delegates a change that must pass review. ChatGPT's Goal mode uses goal text as completion criteria and can use tests or measurements. Cowork's Engineering plugin checks pull requests against requirements, but those checks are optional and do not automatically cover every delivered file.
For a small software team, the difference is where the verification procedure comes from and which assignment it covers. A code patch, weekly report and spreadsheet should each have a clear stopping condition. Compare the published check for the artifact you need, then decide who reviews the evidence before calling the job finished.
ChatGPT connects conversation to longer assignments
ChatGPT supports research, writing and coding, with Projects collecting related chats and sources. Its long-running work documentation describes Goal mode working toward an outcome and completion criteria. You can develop the goal before starting and supply tests or measurements to verify progress.
Dot can delegate to parallel background agents and coordinate them. Enterprise teams can own shared Team Tasks whose runs and results members inspect. Those are useful coordination capabilities, although its public pages do not describe a persistent lead role over a goal, project or team, or a work item individually owned by either a person or agent.
For software work, Codex provides a route from conversation into a repository. Its published pull-request process includes reviewing changes, addressing feedback and preparing changes for a pull request. That matters if your assignment includes responding to a review, not only producing the initial diff.
Cowork keeps planning close to the request
Claude Cowork delegates knowledge work across files and connected tools. Its getting-started instructions describe analyzing the request, creating a plan and splitting complex work into subtasks. You review the approach before execution, and Claude can coordinate parallel sub-agents within a task.
Projects provide instructions, context and memory for future work, and the Productivity plugin maintains a task list that people and Claude can edit. Its public pages do not describe those items as individually owned human-or-agent records with discussion and history. That distinction matters when an assignment moves from agent execution to a teammate's follow-up.
Cowork and chat are being combined across accounts and plans. An account with the unified experience can delegate a larger job in the same conversation as a quick question. Accounts that have not moved still have separate Chat and Cowork options, so do not base a team process on everyone already seeing the same interface.
Completion criteria and review scope need separate tests
Goal mode can use the goal text as completion criteria and accept tests or measurements. ChatGPT also publishes repository-specific review rules for Codex in AGENTS.md. For a code patch, those rules let the review refer to the repository's own expectations.
Cowork's Engineering code-review skill reviews security, performance, correctness and maintainability. With a connected knowledge base, it checks team coding standards and style guides. Connected source control also lets it inspect the pull-request diff, continuous integration status and test results.
The distinction is the assignment boundary. Cowork's checks belong to an optional Engineering plugin, while Goal mode's completion criteria belong to the goal you configure. Neither description means every document or conversation receives the same automatic review, so record the check you intend to invoke for each job.
For the follow-up cycle, ChatGPT's code-review documentation establishes addressing pull-request feedback. Cowork's public plugin pages describe fetching and reviewing a pull request, but do not describe agents revising the branch in response to review feedback. A review-only assignment and a revise-until-accepted assignment should have different acceptance criteria.
A recurring job needs an eligible trigger
ChatGPT's automation docs publish schedules in a chosen time zone and supported Gmail, Slack and GitHub event triggers. Availability depends on plan and workspace settings. One task cannot combine an event trigger with a time-based schedule, so check whether your recurring job requires both entry routes.
Cowork's scheduled tasks run hourly, daily, weekly, on weekdays or on demand. You can review upcoming and past runs. Its public Cowork pages do not describe external events or webhooks starting tasks.
For a weekly repository summary, both publish a scheduled route. If the assignment must begin after supported GitHub activity, ChatGPT documents that event route. Cowork skills preserve procedures and scripts, but its public pages do not describe an enforced multi-step workflow that runs the same way each time.
Permissions and execution location change the brief
ChatGPT's Dot controls specify actions to take, ask about or hand off. OpenAI says these rules are instructions Dot tries to follow and it can make mistakes. They do not replace app permissions or built-in safety requirements, so an instruction to ask is not proof that an output was checked.
Cowork lets you set connector tools to Always allow, Needs approval or Blocked. Manual mode pauses for permission requests inside the task. That gives you a published permission surface alongside the separate review procedure for the result.
ChatGPT's execution modes let Codex tasks use This computer or Cloud, with Codex CLI also working locally. Cowork's Team and Enterprise guide describes cloud and local sessions. Its execution guide says new Pro, Max and Ultra tasks run only in the cloud. Cowork assignments requiring local files, apps or a desktop browser still need Claude Desktop open.
ChatGPT and Claude Cowork, point by point
The table compares answers to the same questions, including source links and the date each answer was checked. Its qualifications distinguish a configured procedure from an automatic check and plan-dependent access from general availability.
Both columns answer the same questions from each company's public pages. Yes means the company publishes the capability the question asks about, Partly means something narrower, No means it says it does not offer it, and Not published means its public pages do not describe it. Every answer links to its source and shows when we last checked it. The oldest check on this page was November 13, 2026.
| What you compare | ChatGPT | Claude Cowork |
|---|---|---|
| Goals |
ChatGPT
Partly
Goal mode works toward an outcome and completion criteria. The reviewed docs do not establish tracking a metric or target against the goal. |
Claude Cowork
Partly
You can give Claude an outcome and schedule a weekly marketing readout with metrics. Its docs do not describe a goal record with an owner tracking a measurable target. |
| Leads |
ChatGPT
Partly
A dot can delegate tasks to parallel background agents and coordinate their work. The reviewed docs do not establish a lead or owner role over a project, goal or team. |
Claude Cowork
Partly
Claude coordinates parallel sub-agents within a task. Its project docs describe sharing and context, not a persistent lead role reviewing and coordinating a project or team. |
| Tasks |
ChatGPT
Partly
Teams can own shared Team Tasks, and members can inspect runs and results. The reviewed docs do not establish work items that both individual people and agents can own. |
Claude Cowork
Partly
Cowork sessions retain progress and results, and the Productivity plugin maintains a task list people and Claude can edit. Its docs do not establish work items owned by either people or agents with their own discussion and history. |
| Recurring work |
ChatGPT
Yes
Scheduled tasks can run in a chosen time zone or on supported Gmail, Slack and GitHub events, with runs available in Scheduled. Event triggers are limited to supported apps rather than arbitrary service events. |
Claude Cowork
Partly
Scheduled tasks run hourly, daily, weekly, on weekdays or on demand, with upcoming and past runs to review. Its Cowork docs do not describe outside events or webhooks starting tasks. |
| Workflows |
ChatGPT
Partly
Skills and plugins provide reusable instructions and tools for scheduled tasks. The reviewed docs do not establish a fixed multi-step process that runs the same way each time. |
Claude Cowork
Partly
Skills save procedures and scripts, and the ticket-triage skill defines ordered categorization and routing steps. Its docs do not establish an enforced multi-step process that runs the same way each time. |
| Autonomy |
ChatGPT
Yes
Dot custom rules specify actions to take without asking, ask about, or hand off. Activity shows requests for decisions. OpenAI says these rules are instructions the dot tries to follow and it can make mistakes. |
Claude Cowork
Yes
Connector tools can be set to Always allow, Needs approval or Blocked. Manual mode pauses for permission requests you allow or deny in the task. |
| Learned autonomy |
ChatGPT
Partly
Dot rules can be edited, and automatic review checks actions against instructions and permissions. The reviewed docs do not describe proposals to change autonomy from approval history. |
Claude Cowork
Not published
Its public Cowork, permissions, memory, plugin and release documentation does not describe proposals to change an agent's authority based on approval history. |
| Memory |
ChatGPT
Yes
ChatGPT and local Codex use separate memories, and a dot keeps notes about ongoing work for later conversations. Local Codex memory files can be inspected, though OpenAI does not recommend editing them as the primary control. |
Claude Cowork
Yes
Cloud Cowork tasks share memory with chat, and projects remember context for future tasks. You can inspect saved memories in Settings and edit or delete individual entries. |
| Planning |
ChatGPT
Yes
In conversation, Codex can propose an approach before editing and develop a goal with measurable success criteria before starting Goal mode. |
Claude Cowork
Yes
Claude analyzes your request, creates a plan and breaks complex work into subtasks. The getting-started instructions tell you to review Claude's approach before letting it run. |
| Intake |
ChatGPT
Yes
The Workspace Agents API accepts external requests and queues them for a published agent, returning a conversation link. |
Claude Cowork
Partly
The Customer Support plugin uses connected ticket history and correspondence, categorizes incoming tickets and recommends a team or queue. Its published workflow describes routing advice, not assigning an incoming request to an owner or agent. |
| Connectors |
ChatGPT
Yes
ChatGPT and Codex support custom Model Context Protocol servers and plugins, including local command-based servers for Codex, and scheduled tasks can be triggered by supported app events. |
Claude Cowork
Partly
Claude offers a connector directory and custom remote Model Context Protocol servers, with plugins for connected apps. Its Cowork docs do not describe webhook or event triggers that start work. |
| Models |
ChatGPT
Partly
You can choose a model and reasoning effort for scheduled work or use defaults. The reviewed docs do not establish automatic model selection per task or step. |
Claude Cowork
Partly
You can select the model for a scheduled task, and the unified experience retains a model picker. Its docs do not describe automatic model selection per task or step. |
| Cloud or your own computer |
ChatGPT
Yes
The ChatGPT desktop app lets you choose This computer or Cloud for Codex tasks. Codex CLI also works against your local repository. |
Claude Cowork
Yes
Team and Enterprise document cloud sessions and local virtual-machine sessions. New Pro, Max, and Ultra tasks run in the cloud only, so the local option depends on your plan and experience. |
| Your own subscription |
ChatGPT
Yes
Codex CLI, the IDE extension and ChatGPT desktop app accept ChatGPT subscription sign-in or an API key for local work. Codex cloud requires ChatGPT sign-in. |
Claude Cowork
Yes
Cowork is included in paid Claude plans and uses the same plan allowance, so an existing Claude subscriber can delegate work without a separate Cowork subscription. |
| Pricing |
ChatGPT
Yes
ChatGPT offers free and paid plans with usage limits and credits for some work. Individual plans cost $0, $8 or $20 per month, with Pro at $100, $200 or $500 per month. Business costs $20 per user monthly when billed annually or $25 monthly. Team Tasks use workspace credits. |
Claude Cowork
Yes
Pro costs $20 monthly or $200 annually. Max costs $100 monthly and Ultra $200 monthly. Enterprise costs $20 per seat monthly, billed annually with a 20-seat minimum, plus usage at API rates. |
| Correctness |
ChatGPT
Yes
Goal mode uses the goal text as completion criteria and can be given tests or measurements to verify progress. |
Claude Cowork
Yes
The Engineering plugin checks a pull request against its stated requirements and can inspect CI status and test results when source control is connected. These are optional plugin checks, rather than automatic checks on every delivered file. |
For software work
| What you compare | ChatGPT | Claude Cowork |
|---|---|---|
| Code review rules |
ChatGPT
Yes
Codex reviews GitHub pull requests using repository-specific review rules in AGENTS.md. |
Claude Cowork
Yes
The Engineering plugin reviews code for security, performance, correctness and maintainability. With a knowledge-base connection, it checks changes against team coding standards and style guides. |
| Adversarial review |
ChatGPT
Yes
Codex can run security review alongside code review and can delegate parallel reviews. The reviewed docs do not establish automatic two-reviewer gating for every high-risk plan. |
Claude Cowork
Not published
Its public Cowork and Engineering plugin docs do not describe an independent second review of risky changes. Published sub-agent coordination and a code-review skill do not establish that review contract. |
| Risk rating |
ChatGPT
Partly
Codex can explain pull request risks and review changed code. The reviewed docs do not establish a rating of the change's merge risk before merge. |
Claude Cowork
Partly
The code-review skill reports issue severity and an Approve, Request Changes or Needs Discussion verdict. Its docs do not describe an overall risk rating for the exact change before merge. |
| Pull requests |
ChatGPT
Yes
Codex can review pull requests, address feedback and prepare changes for a pull request. |
Claude Cowork
Partly
The Engineering plugin accepts a pull-request link, fetches its diff through connected source control and reviews it. Its docs do not establish agents revising the branch in response to pull-request review feedback. |
A fact we cannot find or cannot check is left out, and prices can change after the date shown. Task Machine publishes this comparison and is an interested party. ChatGPT and Claude Cowork are named only to identify them, and there is no relationship between the companies. Has something changed? Tell us.
Who each one suits
Pick ChatGPT if Goal mode criteria, supported app-event triggers and Codex's review-feedback cycle fit your software work. Its custom Dot rules still need supervision, and published reusable skills do not establish a fixed multi-step workflow. Its pricing docs distinguish plan allowances and credits, with Team Tasks using workspace credits.
Pick Claude Cowork if planned file delegation and optional Engineering review fit work you already handle through Claude. Cowork uses the paid Claude plan allowance, and its pricing page lists Pro at $20 monthly or $200 annually. Longer agentic work consumes more capacity than a short chat, and local-session access depends on the plan and experience.
Where Task Machine fits
Task Machine makes requirements part of every Task and checks each one against delivered work before human review. Its Workflow builder adds ordered steps, branches and verifier checks that return failed work with feedback. A person or agent owns the Task, with plan, discussion, decisions and runs on the same record.
Task Machine is the wrong pick if your paid assistant plan already covers the assignments and review procedure you need. For shared ownership and recorded acceptance checks, inspect the ChatGPT comparison and Claude Cowork comparison, then choose one code change and specify the evidence its reviewer must receive.