According to a report by The Decoder on 11 1, OpenAI updated its Help Center article “Understanding ChatGPT and Codex usage limits”: As ChatGPT integrates the agentic features of Work mode into the main interface, starting in 12, Plus, Pro, and Business plans will share a single token-based usage quota between ChatGPT and Codex. More complex models and tasks consume quota faster, while chats using the Instant model do not count against it; remaining quota and reset times can be viewed in settings, and once exhausted, users can wait for a reset, upgrade their plan, or purchase additional credits.
Confidence: Medium (60) · Event Time 2026-11-01
Reasoning
- This marks a shift in ChatGPT subscriptions from “per-message limits / separate limits per model” to a unified token budget: running long tasks in Codex will directly encroach on the same account’s chat quota, and vice versa, requiring heavy users to reassess whether their plan is sufficient.
- The exclusion of Instant conversations means basic chatting remains effectively a flat-rate subscription, but users on lower-tier plans who want to use more powerful models for complex tasks will hit their limits faster; The Decoder suggests that OpenAI’s previous increase in quota for higher-tier plans may have been preparation for this merger.
- The merger of Chat and Work was previously announced; this update finalizes the billing details. Anthropic has also merged Chat and Cowork into a single Claude interface.
Usage
- ChatGPT paid users: Before 12, monitor the new quota and reset time displays in settings, and reserve sufficient quota for chatting when running large tasks in Codex; refer to the Help Center article “Understanding ChatGPT and Codex usage limits” for authoritative rules.
Risks
- OpenAI’s Help Center returned 403 for this section, so the editors did not read the original text; details are sourced from The Decoder’s reporting. Specific quotas for each plan, token conversion ratios for different models, and effective dates are subject to the official pages.
- The statement that “heavy users on lower-tier plans will experience degraded service” is The Decoder’s assessment, not an official statement from OpenAI.