📅 2 November 2026 • TechCrunch
In an email to users, OpenAI said it is merging the Chat and Work tabs in the ChatGPT desktop app into a single experience. In the combined workspace, users can bring files and plugins and have ChatGPT build a website, create a presentation, or act on the web, all in one place. The move ends the split between asking ChatGPT something and putting it to work on a task. The rollout starts with a small test group, so most users will see the change gradually rather than immediately, and Codex remains a separate view inside the app.
Alongside the interface change, an updated OpenAI support page describes a usage change that lands in December: Plus, Pro and Business plans will draw ChatGPT and Codex usage from a single shared token allowance instead of separate budgets. One important detail for planning: chats on Instant models will not count against that allowance, so lightweight conversations stay effectively outside the budget. The flip side OpenAI itself flags in the email: usage draws on the shared allowance by model and task, so advanced models and complex tasks consume it faster, and some users may need to wait for resets more often than today. For business customers this is the difference between tracking two meters and tracking one.
Account management is consolidating too. The remaining balance and reset times move to the Settings, Usage page, giving users one place to check what is left and when it refreshes. Separately, OpenAI has already doubled Work and Codex allowances for the Pro 100, Pro 200 and Pro 500 tiers, effective October 29, so the unification of budgets lands on top of limits that were just raised for the top tiers.
Context: Anthropic made a similar move in September, merging Claude chat and Cowork into one screen. Both vendors are pushing in the same direction within weeks of each other, which makes a single unified assistant surface the emerging default for knowledge work tools: conversation and hands-on work share one context, one toolset and one budget.
IMPACT for practitioners. For teams on Plus, Pro or Business, the December change turns two planning numbers into one: ChatGPT and Codex workloads will draw on a single shared allowance, so capacity planning and internal chargebacks need updating before the switch. Route routine chatter to Instant models, which do not consume the budget, and reserve the allowance for work that uses the stronger models and Codex. When the change reaches your account, check the Settings, Usage page to see the real balance and reset cycle instead of assuming the old split. If you build automations on top of ChatGPT plans, revisit assumptions about separate Codex budgets before December. And once the unified workspace reaches your test ring, re-test your prompts and automations: files, plugins and web actions in one session remove the tab-switching friction between asking and doing, which changes where the practical bottlenecks sit.
Sources: techcrunch.com, bleepingcomputer.com
Continue the conversation: https://t.me/AIAutomationArchitecture · Bluesky
