Like what do you mean gemma 4 and opus 4.6? I don't fully understand ngl. Is it any good?
The specific model is Gemma-4-31B-Claude-4.6-Opus-Reasoning-Distilled on NanoGPT and link: https://nano-gpt.com/models/text/Gemma-4-31B-Claude-4.6-Opus-Reasoning-Distilled
Tbf this is my first week, but I've been using it for coding with high reasoning (and i used a x2 token model) and for RP (also with x2 models) + some general stuff, and I not even close to halfway lmao. If only claude had the same limits, oh boy.
The models I've used the most have been kimi k2.6, deepseek v4 pro and glm 5.1, although I found kimi to be the best of them for some reason. I guess I just didn't test it enough when I was using PAYG
And I'm talking more specifically the more smutty RPs, cus that's the only one I do really. But it's been getting kinda boring doing the same loop over and over now.
I need some tips to either extensions, prompts, or just anything tbh, to make it more fun or to add something weird and totally different. I usually keep replies under 300 words because I'm not really into the novel type stuff, so preferably something that doesn't add that I guess.
Even just a different model would probably help too. Although, I only have 8gb vram (but 32gb ram) incase you have a local model to suggest, so can't run the best ones. Gemma 4 is probably the only one I've been able to run that isn't slow as fuck.
Other than that I have nanogpt sub, but the roleplay/uncensored models there lowkey suck or they're just never available for some reason?
So I think anything would help actually.
