MissingNo: New FREE Hermes Agent Stealth Model
AI News Today | Julian Goldie Podcast
October 13, 2026
MissingNo: New FREE Hermes Agent Stealth Model
Speakers Julian Goldie
Julian Goldie (0:00)
So today, we have a brand new model from Hermes. Agent is a new stealth model. It's called Missing NAR. And I think that means missing number, basically. It's a reference to Pokemon. And this is a new free model you can use inside Hermes Cloud. I'll show you how we set it up in a second and how it performs. And this is free to use. Pretty fast as well, when I've tested it out. We tested it, for example, for the learn command. We also tried building and coding with it as well. I'll show you some examples of what we build with that in a second. And you can plug it straight into your Hermes agentic operating system, like you see right here. So if you're wondering what this is missing though, it is a stealth coding and agentic reasoning model recently made available by News Research. News Research are the creators of Hermes Agent, and you can see the announcement right here. Now it is designed for coding and agentic reasoning.
It's supposed to be highly token efficient. They haven't actually announced who it's come from. Although I was thinking maybe it could have come from News Research themselves because they create a lot of AI models themselves these days. And you can see the details right here. So it's a stealth encoding model. It was announced on the 10th of October.
It's got the Pokemon reference right here and described as highly efficient. It's offered for free on News Portal. And if you want to set it up, you can install it like this.
Or what I like to do is just get Hermes to configure Hermes Cloud for me. And then we can go from there. So you can see that we've already plugged it into our system. What I like to do is I have a separate profile for each Hermes agent model. So for example, we can use step five, which is another free model with Hermes agent. Or we can use missing note, as you can see right here.
And then we can ask it to build something for us, right? So for example, we can plug in a command like this and then start using it inside our agent. One of the things I like about having it inside this system, as you can see, is that we can actually use a voice agent in the meantime. So we have this pop up voice agent, the words with chad, chippity and Hermes. And we can speak to it now. How are you, sir? All right, sir. What are some useful Hermes commands, for example, like forward slash learn that we could use day to day?
SPEAKER_2 (2:11)
That's up tasks before running them. And slash skills helps you find and organize those skills.
Julian Goldie (2:18)
So you can see here that we can basically use Hermes agent as a voice agent whilst we're speaking to Hermes in the middle here. And then also one of the good things about this whole system is like you can switch between three models. So for example, if missing node gets taken down or is not available anymore, or for example, it becomes a paid model or even you run out of tokens, you can switch over to step five or one of the other eight free models on news research as well, which is super useful and powerful here. So you can see an example. Now, I actually did ask it to use the forward slash learn system, and it worked really well.
Worked really nicely for that. So it created a skill name, the path, and it was really fast, I think because it is token efficient, so it was super fast. I did use it for a coding task earlier as well, and it didn't work so well. One thing that I found, for example, is if we test out this game, it's not bad, but it's not like frontier level. If you're using a free model, you can't really expect that anyway, but it's not like gonna replace Opus 5.5 anytime soon or something like that. However, what I do like about it is it's pretty fast at calling agentic tools and also it's free, and you can also use it on Hermes Cloud. Now, the thing I like about Hermes Cloud models is usually they're very quick to respond because they're hosted somewhere else. Let's have a look. So we got this game here. This is the one that we're testing out here, and it did take about five minutes to generate this, but it does actually work as you can see, which is pretty cool.
Obviously, you're not gonna be building ping pong games all day, but you get the point. You can use it that way. Now, we could also use this with custom workflows. So for example, if we go into Hermes bots here, we could use missing null on the side, so we can just create a new one here and add the details from an active profile.
3 more minutes of transcript below
Thousands of transcripts fetched by people building searchable podcast archives
Fetch the whole transcript
The demo key returns a sample episode in full, no card needed:
curl -H "x-api-key: pt_demo" \
https://spoken.md/transcripts/1000651996090Markdown with the speakers named, for your notes, your knowledge base, or anything that makes HTTP calls.
From $0.10 per transcript. No subscription. Credits never expire. Prices exclude VAT, added at checkout for EU customers. Not what you expected? Email us within 14 days with 20 or fewer credits used and we refund the pack in full.
Using your own key:
curl -H "x-api-key: YOUR_KEY" \
https://spoken.md/transcripts/1000796106775