OpenAI merges ChatGPT and Codex into one token pool - what changes for business automation

OpenAI is merging ChatGPT's Chat and Work tabs into one interface and, more importantly for operators, putting ChatGPT and Codex on a shared token allowance pool. If your team runs customer chats on ChatGPT and uses Codex for internal scripts, both now draw from the same budget - watch your quota before it stalls a workflow.

The change was announced in a TechCrunch brief based on an email OpenAI sent to users. Rollout is gradual, starting with select accounts for feedback. Codex stays as a separate section in the desktop app, but the token math is unified. OpenAI says most users won't notice, though heavy accounts may hit resets more often. Anthropic did the same consolidation with Claude's chat and Cowork back in September.

What this means for a small business

If one person on your team is drafting proposals in ChatGPT while another is generating code in Codex on the same plan, you are now competing with yourself for tokens. For small operators this is a budgeting question: either split roles across separate seats, or route heavy code generation to the API where limits are predictable. In most AI automation projects I ship, I keep interactive chat on the consumer plan and push batch work (OCR, document parsing, scheduled agents) to API calls precisely so one noisy user cannot throttle production jobs.

Why it matters: Shared quotas quietly turn two independent tools into one shared bottleneck - audit who uses what on which plan before a Monday morning rate limit stops your pipeline.