OpenAI has merged ChatGPT and Codex into a single unified application, adding a dedicated “Work” mode designed for structured productivity tasks like presentations, tables, data analysis, and professional workflows, according to Quasa. The update follows GPT-5.6’s general availability release earlier this month.
User Growth Hits 8 Million, With Caveats
Tibo, head of ChatGPT and Codex at OpenAI, stated that the combined active user count for Codex and ChatGPT Work could cross 9 million users today, per Quasa. The platform has cleared three milestones in quick succession: 6 million, 7 million, and 8 million active users, with OpenAI resetting weekly usage limits each time Codex hits another million.
The growth numbers come with an asterisk. A significant portion of new users are existing ChatGPT users who were prompted or required to update to the unified app, making it difficult to determine how many represent genuine new Codex adopters, particularly developers, versus forced migrations from the legacy ChatGPT interface.
UX Friction and Missing Features
The consolidation has drawn criticism from users who relied on the previous full-screen ChatGPT experience. The new interface replaces it with a smaller floating window that some users report feels less comfortable for long conversations or extended work sessions, according to Quasa.
Chat history has also become a pain point. Some users report seeing previous conversations in the new window while others say history is missing or incomplete. There is currently no search functionality for retrieving older conversation threads.
$100 Credit Promotion
OpenAI is offering $100 in non-API credits for Codex and Work mode to subscribers paying at least $8/month who share public feedback comparing GPT-5.6 with competing models like Claude and Gemini, per Quasa.
Platform Consolidation as Competitive Strategy
The merge eliminates the context-switching tax between chat and code agent workflows. Rather than maintaining separate products, OpenAI now offers a single surface where users can shift between conversational queries, structured productivity work, and agentic coding sessions without leaving the application.
The move positions OpenAI against Claude Code, GitHub Copilot, and Cursor, all of which already offer integrated chat-and-code environments. The question for developer adoption is whether the Work mode’s structured output capabilities (tables, slide outlines, project plans) can differentiate enough against competitors that have been iterating on coding workflows for longer.
Early user feedback on Work mode has been mixed but largely positive among those who have tried it, with reports of improved handling of structured outputs and better context retention across longer sessions, though careful prompting remains necessary for consistent results.