Google has a growing Gemini problem that readers find hard to ignore
More and more users are calling out Gemini's apparent decline.
Google's Gemini artificial intelligence system is facing criticism as users report that its performance and quality have declined in recent months. Reddit threads and Google's own polling indicate that over 60% of respondents believe the AI model has gotten "dumber" or "stupider," with nearly two-thirds of those users also expressing frustration with the service.
In a recent poll, 65% of respondents reported experiencing a decline, while 20% said they did not notice any change. Commenters on forums have noted that Gemini's responses are less comprehensive and that the tool is now configured to use less compute power, resulting in reduced performance. Some users have even switched to competing AI services, such as Claude Code, finding them to be vastly superior.
While some users still find Gemini useful, a significant portion of the user base appears dissatisfied with its current capabilities.
Written by urgent.news from Android Authority's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
- Google reveals new Gemini 4 Carbon AI model, narrowly beating Claude Opus 5.5 — frontier AI model is also the least likely to hallucinate, benchmarks claim tomshardware.com
- After testing it on 100 photos, Google Photos' new Gemini integration is both good and horrible androidpolice.com
- Google may already be having second thoughts about this recent Gemini change androidauthority.com
- Local Government Elections: I turned to Google Gemini for advice, and this is what it told me iol.co.za