@mattshumer_ Not that i run it locally but GLM 5.3 Flash from Fireworks is far from useless! This is my default model in Hermes and runs all day 24/7. I have no need for a big expensive model for this work!
@MiaAI_lab I dont know about benchmarks but as a default model in Hermes Agent GLM 5.3 Flash is exceptional! It has not let me down since running it.
For a everyday smart assistant this is the one - its perfect π
You do not need a big frontier model anymore.
Sending all my receipts and invoices to Hermes via telegram or email is fast and seamless, it files everything and can give me a quick expense report whenever I ask. This help alone is worth paying $14 per month for a VPS!
Setting my default model in Hermes to Luna Max has made a huge improvement to my codex usage. I find it can be a little slow at times but it does get the job done. Luna is an excellent model π
@thatguybg I use my hermes agent everyday and night, its always upto
Something. I have visited the web site maybe twiceβ¦ who the cares what it looks like πππ
@sattyyouneed Making sure the plan looks rock solid on paper before starting work with codex.
Iterate intensively with pencil, paper and gemini or chat gpt first.
Keeps the brain π§ humming as well!
@bridgemindai Run Hermes on a VPS with your Sub / API of choiceβ¦ itβs brilliant!
No dusty mac minis or $420 per month Grok subscription required thanks π
@tansawit@grok Just ask Hermes to set it up as you have described. Itβll get it done and it works well!
I have my default agent send work off to several agents / profiles and then it comes back to me with the result. I never message the other agents / profiles directly.
Iβve decided to remove all the other shiny distractions and stick with the GPT / Codex sub for my daily go to.
Grok is excellent but just way out of reach for the average user. Iβve removed the app.