
Model Problems #12: Hermes 3 70B
Honestly, I forgot about this model. I don't even have an opinion on it, but i do remember when there were some people being fans of both Hermes models, so. Any thoughts?

Honestly, I forgot about this model. I don't even have an opinion on it, but i do remember when there were some people being fans of both Hermes models, so. Any thoughts?
Signature Classic. Wayfarer large was a go-to fallback for deepseek last year personally. It carried the adventure side of stories fairly well, but it's definitely overshadowed by the newer models, I honestly hope a WL 2 is underway.
Thoughts?
We begin with the large models!
DynL is pretty good in my opinion for allowing max 32k context, and for the past year i used to fall back on it once the current 16k model cannot fit all memories referenced in a long adventure. My issue will always remain that we don't know what the model switches to, sometimes we'll get good output, other times it'll be mediocre.
Any thoughts?
credits to Pod in the bag-xml discord server we won bois now i can use "Log in with facebook" on every old ass game on ios 6
Sorry I wasn't able to post one yesterday, had a tough day. Now i gotta go through the older post's replies.
Anyways. what a throwback! Mistral Small. I used to use it originally and mained it for a while until Dynamic DS and Raven came about, now I forgot about it's issues since it's a deprecated model, so i have 0 valid opinions here.
Any thoughts from any (former) MS/3 users?
V4F was pretty good in my opinion. It carries character interactions and personalities so well, and even has a good grasp on cultural references. I've used this as my main for a bit, especially since it's the first DS model ever released for free tiers, any thoughts?
New free model.
Tells alot of fables, doesn’t it? Thoughts?
Hearthfire was something i mained for a bit when it came out. I’d say it’s fairly good for its realism, but nowadays it’s been overshadowed by other models. I haven’t used it recently.
Thoughts?
Haven’t messed with it much but it has lots of ironic tendencies. Any thoughts?
Most complain about WS2’s often forced scenario aggression element. The model is explicitly designed for that rather than WS1’s usual challenging difficulty.
I personally think WS2’s should adapt a new approach that analyzes & adjusts to the tone, mood, atmosphere but retains the challenging element WS1 had. In some kind of way where there’s potential of something to go really wrong really quickly, making choices really matter.
What do you all think?
For the past year i seen majority opinions change from “Muse is bad” to “Muse is good.” For me, I think it really depends.
What are ways we can make it better?
Thoughts?