Maybe it’s just me, but the model seems to have gotten significantly since today. Has anyone noticed a decline in the quality of its responses today?
2 Likes
Yes, it’s becoming worse. The last 1-2 days have been annoying.
The models (both 2.5 and 3) are extremely sticky brained now, and just had it forget completely, that it was a code agent. Couldn’t do anything with the files because it didn’t have access. Then it implemented some crappy changes anyway.
2 Likes
When 3.0 came out it was amazing for a week or three. Since then, it’s been a steady march towards being unusable. I have to think Google really cranked up the generosity with the model release to “catch up” to Open AI, and it definitely made some headlines, but now that usage has increased, they have really throttled their models to cut costs. Performance in AI studio, the Gemini web interface, and Antigravity, are all just awful now, 2.0 level quality or worse.
I’ll likely cancel my subscription once this month is up. Given the ability to use 3.0 preview in AI studio, there’s really no reason to be paying for something that has become a worse version of the product it replaced. I should probably migrate over to Claude honestly. I enjoy the Google ecosystem and Gemini integration, but this just isn’t worth a subscription any more.
1 Like