OpenAI introduces free banked reset system for Codex users

OpenAI has rolled out a “banked reset” system for its Codex coding assistant, giving subscribers the ability to accumulate rate limit resets and deploy them whenever they actually need them. The feature is free for eligible users on Plus and Pro plans, marking a notable shift from the old model where resets happened on a fixed schedule whether you needed them or not.

The mechanics are straightforward. Instead of resets firing automatically on a predetermined timer, eligible Codex users can now save their resets and use them on their own schedule. Users can track their available resets through the profile or usage menu inside the platform.

OpenAI has paired the launch with a limited-time referral program. When an existing user invites a friend and that friend completes their first Codex interaction, both the referrer and the new user earn an additional reset. Each eligible user can refer up to three friends, and the program runs for a two-week window.

No paid upgrade tiers or premium banking options exist for this feature. It is entirely free for users already subscribed to qualifying ChatGPT plans.

From April through June 2026, user complaints about unpredictable and inflexible reset schedules became increasingly vocal across developer communities. Codex has grown significantly, crossing the 3 million weekly active users milestone and continuing to scale from there.

The referral component offers resets as a referral incentive instead of discounts or credits. Getting a new user to complete their first Codex interaction is the conversion event that matters.

Codex competes in an increasingly crowded space where GitHub Copilot, Anthropic’s Claude, Google’s Gemini Code Assist, and a growing roster of open-source alternatives are all fighting for developer mindshare.

There is no token, no blockchain integration, and no digital asset play embedded in the banked reset feature. OpenAI remains squarely focused on improving the usability of its coding tools.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.