New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.


Comments
QNow4
Apparently it's disappointing
Qwen4 Max is pure ass
this ^
@Neoon you were a qwen3.8 flash next stan right?
Me too I actually like qwen3.8 flash next so much more than deepseek it’s much better
More than dsv4.1 flash?
Yes but I dont mention it in front of @wadhah because he gets angry
is that because he likes to rp to deepseek
deepseek is lowendgirlfriend if it was good
nothing comes close to ds for rp
I tried qoduck (which people say was qwen 4) while it was available, but I was not really blown by it. Deepseek 4.1 flash 1017 is still the goat for me.
And @Rubben sure does know ass
i mostly use free GLM 5.3 Flash (z.ai gives 300mil tokens per day free, and 600mil on weekends) and Nvidias free Nemotron models (Lightning for error analysis and simple tasks, Super for basic code and Ultra for research), and Googles free Gemma as research backup with Gemini 2.5 flash lite for translations (if Nvidias translate model returns overload/429, Google is also free) , for frontier tasks i use Meta Muse Spark 1.3 Contributor at 10/20c per Million tokens.
I see no reason to pay over a dollar per million if you know how to avoid sending secrets to training enabled endpoints (and eg. Antrhopic will train on your input anyway). For compliance things i use Mistral via Openrouter EU (which is allowed as it is inside EU, no need for direct Mistral endpoint in France).
I've been using Minimax M3.1 Flash for general stuff, its really impressed me
i think qwen4 max is just a cheaper/faster k3.1 replacement but nothing that blows it out of the water
is q4 max is this mid i have low hopes for 27b
disappointing
Our live support by Frankie system is based on Gemma 4, which takes up most of our 9999 plan.
Does QWhen4 use less RAM?
QWhen4 will take up less RAM and will run faster than Gemma. Gemma is slow and annoying.
However, you will not be able to ask your Live Support by Frankie anything about events in 1989