ChatGPT to merge Chat and Work, with ChatGPT and Codex sharing one token budget from December and paid accounts capped per device and card
The Decoder reported on 1 November, citing updated OpenAI Help Center pages, that ChatGPT is folding its Chat and Work modes into one interface; from December, Plus, Pro and Business plans will share a single token-based budget across ChatGPT and Codex, while chats with Instant models will not count. OpenAI is also adding device and payment-card limits to individual paid accounts from Go to Pro 500.
According to the support page as described by The Decoder, the budget is counted in tokens, so more complex models and heavier tasks use it faster. Users can see their remaining balance and reset times in Settings and, once it runs out, wait for a reset, upgrade or buy extra credits. Ordinary chats with Instant models do not draw on it, so ChatGPT remains a flat-rate service on paper, but The Decoder expects heavy users on cheaper plans to get lower-quality results, which is the outlet’s inference rather than OpenAI’s statement.
The merger responds to ChatGPT’s growing complexity: The Decoder says users currently choose between two modes and 11 model settings. OpenAI had already announced that Chat and Work would merge and recently doubled Work and Codex limits on several Pro tiers. Anthropic has already combined Chat and Cowork in Claude.
A separate Help Center update adds account limits. Users who link too many subscriptions to one payment card or device cannot start new ones, and the number of paid accounts per device is capped; at the cap ChatGPT shows “Device limit reached” and the affected account can no longer be used on that device, and logging out does not free a slot. The limits cover ChatGPT Go, Plus, Pro 100, Pro 200 and Pro 500; Free accounts and Business, Enterprise and Edu plans are excluded. OpenAI frames this as a security measure and points users to extra credits or usage resets; The Decoder expects it to hit people rotating several cheap subscriptions to stretch their usage.
The Decoder’s reports give no budget sizes, no exact December start date and no per-plan device caps. Direct retrieval of the OpenAI Help Center pages was refused (HTTP 403), so these details rest on The Decoder’s account.