DeepSeek v4 (both Flash and Pro) is kinda slow and mostly returns only error messages on @NousResearch Plus subscription on Hermes Agent. This happens especially around midnight-early morning (EST).
Using GPT models on the same Nous subscription is faster and more stable, but more expensive.
Is this expected or did I configure my instance incorrectly?
@jazzychad Ya this is so dumb, in my country even our local TV uses the term “Happy Football” for the event, and they are literally the official broadcaster
It actually makes the marketing very hard and many people don’t know that the World Cup is currently happening