I am writing to express my extreme frustration regarding the recent silent downgrade in image generation quality. Up until a couple of days ago, I was consistently able to generate high-quality 2K images. Now, suddenly, everything is locked at a blurry, useless 1024x1024 resolution.I pay for a premium subscription to get premium features, not to have essential tools suddenly nerfed without warning. This new limitation completely breaks my workflow. I use these images for further animation and content creation, and these 1K generations simply fall apart and turn into a pixelated mess when put into production.I am paying for high-fidelity generation, not this downgraded restriction. I didn't pay for this. If the high-resolution (2K) generation capabilities are not restored immediately, I will be canceling my subscription and taking my money elsewhere.Bring back the normal quality.

Initially, everything was fine, but over time the app just kept getting worse. It can’t properly work with photos and always says it can’t edit them, even though this issue seems to happen only to me. It also quickly forgets the context of the conversation.The new interface is much less convenient. Before, editing one of its messages only took one tap, but now it requires several extra steps, and it’s easy to tap the wrong button by mistake.The fact that it can no longer work with photos was especially disappointing because that’s the main reason I downloaded the app. Unfortunately, with every update something seems to get worse—whether it’s the interface, the memory, or just constant lag and bugs.

I actually like using Gemini overall — it’s smart, helpful, and has a lot of potential. But the voice recording feature is an absolute disaster and makes the app almost unusable for any serious work.When you try to speak into the microphone — whether asking a question, dictating text, or recording someone else — it constantly cuts off after just 2–5 seconds by itself. You have to start over again and again. This happens way too often.It’s especially infuriating during important moments: conferences, meetings, interviews, or when you’re trying to quickly capture someone’s question or idea. Instead of saving time, this “feature” wastes a huge amount of it and kills your flow. I can’t believe Google released and still hasn’t fixed such a basic and critical bug in their flagship AI assistant.Please fix this voice recording issue as soon as possible. It’s a deal-breaker for productivity. Until then, 2 stars at best.

Please fix the bug regarding outdated information. We were talking about iOS, and I brought up the topic of iOS 26, and he started claiming it doesn't exist and that it will be released around 2032, and the current maximum is iOS 19. But if I switch to the Pro model, he says it exists. And if I then ask again in the same chat without the Pro model, "So does it exist???" He replies, "No, it doesn't exist." And if I tell him that he just wrote that it exists, he'll say I lied or that it glitched and apologize. But if I switch back to the Pro model, he says iOS 26 exists (I hope my words are read). Thank you.

I am writing to express my deep frustration with the app's voice recognition functionality, which is currently completely unusable. I have paid for a full annual subscription, and right now, I feel like I am completely throwing my money down the drain. I honestly do not understand what the developers are being paid for, considering the app receives updates twice a week, yet this core feature remains fundamentally broken.For me, using voice dictation is an absolute necessity, as I do not have the ability to type manually. I completely rely on this feature for daily use, which makes these constant technical failures even more unacceptable.Here are the specific, ongoing issues: Frequent Crashes and Gibberish: At least 5 out of 10 voice inputs fail completely, rendering as useless text like "1 2 3 MMM AAA". It never works reliably. Word Truncation and Hallucination: The system only captures a small fraction of a spoken word and then tries to guess or invent the rest, completely distorting the original message. The Worst Performance on the Market: While the AI agent itself can sometimes be decent, the speech-to-text integration is the most pathetic and frustrating on the market. For comparison, ChatGPT captures audio word-for-word and sound-for-sound exactly as dictated. In this app, I get either wild fabrications or error-driven gibberish.It is completely unacceptable for a paid service that pushes updates twice a week to fail so miserably at basic dictation. I expect this to be escalated to the technical team and fixed immediately, as I am currently paying for a broken product.

The latest update completely broke the Tools menu navigation on iOS. When I tap the tools button, the menu doesn't expand—it gets stuck as a tiny collapsed bar at the bottom of the screen.I have to tap it a second time, and then it explodes into a full-screen overlay that covers everything. What used to be a single tap is now a three-step mess: tap, tap again, then manually resize or close the window. This is a massive UX regression. Please bring back the adaptive bottom sheet!