saiga-nemo-12b / qwen3.5-9b-claude-4.6-9b
a fine-tuned Qwen 3.5 model trained on Opus 4.6 data and reasoning came out 4 months ago, but i only noticed it now :):
9b parameters, and it can run on just 5 gb of vram.
first impressions are very nice.
first impressions are very nice.
i suggest we decide whether it deserves to replace the current default model. please vote if you’ve tried both models.
if yr hardware is powerful, try q5 or higher:
which model is more fcking awesome?
saiga_nemo_12b.Q4_K_M
7 votes
Qwen3.5-9B-Claude-4.6-HighIQ-THINKING-HERETIC-UNCENSORED.Q4_K_M
2 votes
9 users voted
local model
saiga nemo
qwen
claude
uncensored
Berdy Berdy
Hi, just a word to share my impressions: on one hand, this Qwen model is censored; on the other, it fails to capture intensity and passion as well as Saiga Nemo—a model whose Q4_K_M version, in my opinion, performs better when it comes to telling romantic stories. As for the other modes, like Dungeon, I don't know.
Jul 02 14:54 (changed)
1
ASLPro3D ArtSimulatesLife
Honestly, I use "ChatWaifu_v1.4.Q5_K_M.gguf" which is far less censored (but you have to be careful and set rules and limits with it) and handles storytelling and eroticism much better than "Saiga Nemo"... which I think is a better default for new/inexperienced users than "Qwen3.5-9B-Claude" is... as my test of the Claude is a bit restrictive and not sure new players coming on to play an "erotic game" will get as good of an experience as the will with Saiga Nemo... but that is coming from a player that uses a much more... erm... erotic LLM for you game.
Jul 04 05:53
Misaka 10777
Of the two models here, I ran Q5 of both for a while in Sandbox and main story. It didn't take long for Qwen to start looping into itself and demand a scenario to end, meanwhile Saiga permitted SIGNIFICANTLY more freedom.
Jul 04 21:49
