chatgpt.com flag changelog (20260624-010039Z → 20260624-070058Z)
Readable flags
- Upload / rate-limit wording experiments cleaned up (7 flags removed): attach_accepted_files_on_upload_limit, per_upload_limit_cta_copy_variant, per_upload_limit_wording_variant, rate_limit_exhausted_toast_intent, show_per_upload_limit_wording, show_rate_limit_exhausted_wording.
- Google One Tap experiment cleaned up: should_disable_google_one_tap_on_ad_like_mweb_landings removed.
- New flag added: show_recommended_treatment — gates display of a recommended treatment / option.
Models
- gpt-5-*-auto variants removed (gpt-5-1-auto through gpt-5-auto).
Statsig
- 1 gate, 2 layers added; 2 gates, 3 layers removed (hashed).
- SDK updated 0.19.1 → 0.19.4.
No official release-note entries supplied.