>>544738479
How am I not making sense tho? Their fine-tune hasn't been out yet for 4.6, so the glm you've been running thus far has just been the base model. We also don't know what quant they are running. Ideally Q8, however they don't state. The privacy policy is not a NovelAI exclusive feature either, many providers provide the same privacy protection. Also, a context size complaint is not meaningless, if you say 36K is decent, then you haven't tried anything above that, it's genuinely a case of, once you've experienced you wont go back. 36k, is genuinely archaic for the price they are asking for at that tier.
As I stated before, if you like NovelAI, you like it's convenience, sure stick with it, were they to let me import any outside API into their UI, I would still be using NovelAI just for that as I love their UI. But to say NovelAI in it's current offering is "fine" is absurd. Also NanoGPT offers more than GLM, they are access providers, so they provide you with whatever you want, Deepseek, GLM, Claude bunch of Finetunes that people requested, And none of that is a NanoGPT exclusive thing either, you can choose whatever API provider you want.