OctoLink GEO

Codex Merged into ChatGPT: OpenAI’s Strategic Shift from Subscription to Usage-Based AI

Author Editor
Codex Merged into ChatGPT: OpenAI’s Strategic Shift from Subscription to Usage-Based AI

OpenAI’s Codex, once a standalone coding tool, is now integrated into ChatGPT—reflecting user behavior shifts and a pivot to a usage-based pricing model...

OpenAI Codex ChatGPT GPT-5.6 AI Pricing Usage-Based Billing SaaS Industry

OpenAI’s recent launch of GPT-5.6 grabbed headlines, but the more significant development is the integration of Codex into the ChatGPT desktop app as of July 10. The three modes—Chat, Work, and Codex—are now unified under one interface, accessible even to free users. While media framed this as a 'unified entry point,' OpenAI clarified Codex’s capabilities are now built into ChatGPT.

User data reveals an unexpected trend: Codex has over 5 million weekly active users, with more than 1 million non-coders repurposing it for office tasks like report writing, data analysis, and project planning. This use case emerged organically, driven by users rather than product design.

The merge followed organizational changes: in May, OpenAI merged its ChatGPT consumer team with the Codex team under co-founder Greg Brockman, with product integration coming two months later. The UI now prioritizes Work and Codex modes, while the classic Chat tab is tucked into a sidebar—old ChatGPT is rebranded as 'ChatGPT Classic.'

Monetization is at the heart of this shift. OpenAI moved from subscription-only to a 'water meter' model: free app access, but usage-based billing for Work and Codex tasks. GPT-5.6’s three tiers (Sol, Terra, Luna) have tiered pricing—Sol at $5/$30 per million tokens (input/output), Terra at half that, and Luna at $1/$6. Subscriptions now determine usage pool size, with overages requiring additional purchases.

This model impacts the SaaS industry: the 'SaaS Doomsday Theory' suggests AI agents could replace traditional subscriptions. ServiceNow’s stock dropped 62% from its high, though recent rebounds show market uncertainty. Gartner estimates AI agents could divert $234 billion in enterprise software spending by 2030, and incumbents like ServiceNow now derive half their new contracts from usage-linked pricing.

Sources

Related reading