ChatGPT iOS UI is a complete mess for me. Mixed old and new “Liquid Glass” interface
Users report a chaotic mix of old and new interface elements, with reinstalls only temporarily fixing the issue.
OpenAI's rollout of its new 'Liquid Glass' design for the ChatGPT iOS app appears to be plagued by a significant bug, leaving users with a jarringly inconsistent interface. According to user reports on Reddit, the app displays a chaotic mix of old and new UI elements simultaneously. Normal text chats stubbornly retain the old interface, while sections for Images and group chats show the newer, sleeker Liquid Glass design. Compounding the confusion, the app's left sidebar has also updated to an even newer layout, creating a visual patchwork that suggests different parts of the app are running different versions of the codebase.
The bug exhibits a particularly frustrating pattern. Users have found that completely deleting and reinstalling the app results in the correct, full Liquid Glass UI appearing upon login. However, this fix is entirely temporary. As soon as the app is closed and reopened, it reverts to the broken, mixed-state interface. This has led to speculation that the staged rollout mechanism—a common practice for gradually releasing new features—is malfunctioning. Despite users contacting OpenAI support for months, there appears to be no official acknowledgment or fix, with some support agents reportedly unaware of the 'Liquid Glass' interface mentioned in OpenAI's own marketing materials. The persistence of this visual bug undermines the user experience for a flagship product from a leading AI company.
- The ChatGPT iOS app displays a broken mix of old UI and the new 'Liquid Glass' design simultaneously.
- A full app reinstall only temporarily fixes the issue, with the mixed UI returning upon relaunch.
- Users have reported the bug to OpenAI support for months without a resolution, pointing to a flawed feature rollout.
Why It Matters
A buggy UI rollout for a major app damages user trust and highlights potential quality control issues in deployment pipelines.