Codex and ChatGPT Work limits reset after high-compute settings drained usage, with a second reset promised that day.

A confirmed provider-initiated counter clear for the audience below.

What happened

Provider / product: OpenAI / Codex
Plans: All applicable plans
Announced: Jul 10, 2026, 17:59 UTC
Cleared by: Jul 10, 2026, 17:59 UTC
Audience: all users · Codex and ChatGPT Work users
Model or pool: Not specified
Provider-stated reason: Launch feedback: high-compute settings drained limits without clear signalling

Primary evidence and context

Follow the original posts or documentation. Announcement posts and completion evidence can say different things; check each source's claim.

Source 1: @thsottiaux ↗staff post · confirms completion

Research notes

Usage limits were reset, with a second reset promised later that day. Defaults and the model picker were also changed to steer users away from the most expensive settings.

Discovery credits

These feeds helped us find the event. The original provider sources above are the evidence used for this record.

Correction log

Sep 22, 2026, 22:17 UTCeffective_at: None -> 2026-07-10T17:59:43Z · 'We have reset usage limits across Codex and ChatGPT Work'.