19
Gemini 3.5 Flash Lite updated?
Log in to comment.
- @user_11401· 1mo ago
I don't think the model itself is an awful downgrade in every respect. My chat, which I regular on, is still going pretty strong, even with the slightly slower responses. I won't lie, though, having it sometimes failing to generate a response or claiming it's unavailable for the moment is genuinely pmo when I'm otherwise extremely immersed in what's happening
- @Zamasusappleapple· 1mo ago
time to let out the beast, big harp 🐺
@Slut_Queen· 1mo agoGemma Model definitely is slow compared to the GL 3.1 Flash Lite, but so far i haven't come across any quality issues. I do agree that waiting so long for a reply is rather icky, but if it gives me BETTER quality than GL, then i can compromise with the speedy message generation.
Fair. I understand everyone’s point. It is working a little better now, so I suppose that they have fixed it. I feel frustrated about it, still, but there must be change, no?
- @Slut_Queen· 1mo ago
@BlueIceRain actually i kinda take it back. I think it is still fluctuating, because it took yet another 5 minutes to generate a message worth 400 tokens(?) even though i have set it to full 1200 tokens. GL never did that mistake.. so i guess they fixed the speed issue, but dropped off a little on the quality.
Yes, this is exactly what has been happening. My max tokens are set to 2040 (estimated as I have not looked). Yet I get things that are worth less. I saw a post similar to this one that told someone ‘not to waste their money or time’ (quoted DIRECTLY from post) to someone that was new on the platform.
- @user_19392· 1mo ago
its not even responding its just taking forever
Then it had the audacity to say ‘unable to load’. Like, really, we are valid in this frustration…
- @user_5f2afebe· 1mo ago
I hate it. C.ai already did this, change all the fairly good free models for one single bad free model. If harpy follows the same path, I won't waste my time in here. I chose harpy because the characters sounded fairly real and compelling even using the free models, but now the characters have lost all their dept because of this 'Gemma' trash model. If they want to follow C.ai path, I won't be here to see it.
Agreed, but I just went back on C.Ai on my burner account and they make you WATCH AN AD just to chat. Now, whenever I try and generate a message on this website, it just comes up with the ‘couldn’t generate response’ error. This website/app is officially the worst. They should just go back to the original models!
@Vaeluneth· 1mo agoi don't like the model
- @Vamalis· 1mo ago
I do not think it's the model really, maybe they just need time to fine tune/code it. I've used it on another site with fairly high success for a cheap option. The other site has no free options and it costs per message with credits, and their "budget" model uses gemma 4 26B. i have no big issues with it there, but here testing it was pretty bad so far. But i have seen it work, and work well. I am not sure if i can say the name of the other site on here in a positive way. I know cai would delete the post if I mentioned another in a positive way.
True, I understand it. It works perfectly fine, perhaps I was over exaggerating a little…
- @user_48e6c297· 1mo ago
Bro I’m just going to stop using it. I’m not waiting 7 years for one damn responds. So annoying. Like I get that they are trying to upgrade or whatever but this honestly has made it MUCH worse. Way to slow
@Thomasandre· 1mo agoit's SLOW AF, truly. Took 47 years to think just to get "Failed to generate"
@Zeeeee· 1mo agoEXACTLY what happened to me. I agree with you.
I know, even my own brother agrees with me.
- @Zeeeee· 1mo ago
Ngl I HATED it. Too slow, nothing better or worse than Gemini 3.5 Flash Lite.
And now it is the only free option, I might as well go back to C.ai, it actually has a quicker loading time.