ChatGPT Work is OpenAI's agentic workspace, launched on July 9, 2026 on top of the GPT-5.6 model family (Sol, Luna, and Terra). It is built for knowledge workers, founders, and analysts who need a system that takes an outcome — a website, a report, a spreadsheet, a slide deck — and executes it across their connected apps and files over hours, rather than just answering a question.
Core Features
- Long-running background execution: assign a multi-step outcome and ChatGPT Work plans, splits, and runs it while you step away.
- Connected-app context: it pulls from Google Drive, Slack, and GitHub so output reflects your real data instead of generic filler.
- Finished deliverables: it ships working websites, research reports, spreadsheets, and presentations, then iterates on your feedback.
- Usage-based metering borrowed from the Codex model instead of a flat chat subscription.
- Cross-surface availability on web, iOS, Android, and desktop (Mac and Windows).
Use Cases
- Founders drafting an investor update deck and a supporting financial model in a single pass.
- Analysts turning messy CSV exports into a cleaned dataset plus a written briefing.
- Operations teams generating a client-ready proposal from scattered email threads and docs.
Pricing
ChatGPT Work is available to ChatGPT Plus ($20/month), Pro ($200/month), Team, and Enterprise subscribers, with usage-based billing for long agent runs similar to Codex. Access and model choice depend on plan, workspace, and region; business-workspace controls differ by tier.
Our Take
Best for people who want a general-purpose 'give me the result' agent rather than a chat buddy. The trade-off is breadth: because it owns no fixed workflow, you must state the outcome clearly or it will wander. Compare it with focused tools in our efficiency-improvement category at https://aifreetool.site/tool-category/efficiency-improvement/.
