AI News Today
← All episodes
Episode 1 · May 13, 2026 · 05:08

Run Hermes Agent FREE With Step 3.5 Flash 🤯

Use Hermes Agent for FREE with Step 3.5 Flash (15 Days) — Full Setup GuideThe script demonstrates how to set up Hermes Agent to run for free using Step via the News Portal configuration, highlighting that Step 3.5 Flash is available free for the next 15 days. It walks through installing Hermes from GitHub, running `hermes setup` in the terminal, selecting News Portal, switching to Step Fun/Step 3.5 Flash, skipping remaining configuration prompts, and restarting the gateway to confirm the model is operational. It explains what Step 3.5 Flash is (262K context window, ranked 16th on OpenRouter Weekly), notes it’s strong at reasoning but weaker at coding, and mentions other free options like Qwen 3.6 Plus and OwlAlpha (1M token context on OpenRouter). It also shows how to switch models with `/model` and promotes the AI Profit Boardroom community and training.00:00 Free Hermes Setup00:28 Terminal Install Steps01:05 Verify Step Flash Works01:21 Model Specs Benchmarks02:04 More Free Model Options02:41 Install Switch Models03:46 Training Community Pitch04:19 Live Q&A Wrap Up

Full transcript

So today I'm going to show you how to set up Hermes agent for free with Step. Right. So you can see here, Step Fun has said 10 days wasn't enough. Step 3.5 flash is back on this research portal and it's free for the next 15 days, which means that you can use Hermes agent for free over the next 15 days, which is pretty awesome, right?

So you can use one of the most powerful agents in the world for free using Step. As you can see now, how do you set this up? So what you want to do is you go to terminal like this, and then from here, you're going to run Hermes setup, right? And inside Hermes setup, you want to make sure that you have news portal set up as you can see right here, right?

So make sure you have news portal set up and you actually see this two free models available, right? So you can use Quen 3.6 plus, and you can also now use Step Fun with Step 3.5 flash for free, which is what we're going to switch to now, right? Then you can just press enter to skip through the rest of the configurations, restart the gateway to get started. And once we're waiting for the new gateway to get set up, let me just set up Hermes here.

All right. You can see, we now have Step 3.5 flash here, right? And if we type in working question mark, just to make sure it's working, it says, yes, I'm here and operational ready when you are. And so we can use Hermes agent with a free API called Step 3.5 flash, and it's pretty powerful and easy to use.

Now, if you're wondering, okay, what is Step 3.5 flash? How does it compare to everything else? So this is Step 3.5 flash, as you can see. And essentially this is a 262 K context window.

It's ranked 16th on OpenRouter weekly. And the most commonly used apps that it's used with are Kilo Code, Hermes agent, and OpenCore code as well as Lemonade, right? And so you can see the benchmarks here in terms of how it performs. It's pretty good at reasoning.

It's not that great at coding. So you can see a coding. It doesn't perform so well, but at reasoning is very good. And again, like normally this is a paid model, so you can use it for free, which is great, and you can use it for free for 15 days and then bear in mind, like there's always new models coming out for free as well.

So for example, yesterday, Quen 3.6 plus came out for free with Hermes agent. Step 3.5 flash is now free. And then also there's another model, this one final model, which is our alpha. And that's a million token context window and available for free with OpenRouter.

So that's three different ways to use Hermes agent for free, but I would try to step on. The other thing that I would say here is like, it seems to be really fast replying as well, which is great. So if it speeds up your Hermes agent and you can use it for free, that's great. And then it also gives you the context window count here as well in terms of tokens, which is pretty awesome as well.

So if you're wondering, like, how do you set up Hermes, to set up Hermes, you just follow the instructions on the GitHub here, and then you can get the setup. Right. So you just literally copy and paste this command. You go into your terminal, you paste that in and you hit enter.

That's how it's set up. Also runs on windows native as well. Right now, if you're wondering how to set up news, how to set up step fun. So you just type in Hermes setup inside the terminal.

Once you've installed it, then you're going to select news portal, and then you just make sure that you got step fun 3.5 flash setup, as you can see right there. Right. Then once you've done that, you just need to restart the gateway with Hermes inside your terminal, and you'll see the model come up as that. Right.

If you ever want to switch your models inside Hermes at any point, you can type forward slash model like this type forward slash model, and then you can switch between them. So you can switch between news portal, open router, anthropic, whatever you've set up. Right. Even LM studio or Lama, if you set that up previously, but obviously we're focusing on news portal today with step.

So that's exactly how to use it, how to get it for free, super easy to set up. If you want more training on Hermes agents, how to use them, et cetera, you can get that all inside the AI profit boardroom link in the comments description, or go to the AIprofitboardroom.com inside the community. You can ask questions, get help and support whenever you need to inside the classroom, you can learn how to use Hermes agent. We actually have loads of training on this sort of stuff with new daily guides, new daily updates, et cetera.

Inside the calendar, you can jump on weekly coaching calls. We have four per week inside the map. You can connect with people in your area using agents like you, and that's all inside this community. So feel free to get it link in the comments, description, or go to the AI questions we've got here.

If I was running a local setup, yeah, I would just ask, I don't use PC, right. And I don't use local models that much. So I think the best thing you can do, if you want to figure out how to use local models or what's the best setup, just ask Claude, paste in the latest information on our Lama, the requirements on the latest models, et cetera, and then ask Claude, and it will give you some ideas. Are you real or AI?

I am real, my friend. Is the Hermes native desktop version still working? Yes. So you can see it working right here.

Here's Hermes agent. We actually used it earlier today and it's working perfectly fine. Top notch. Thanks.

I think Hermes self-created the business version of itself. That's pretty crazy. I never heard of that, but it sounds interesting. It created the roles and jobs.

That's nice. Nice. Yeah. It's pretty autonomous.

More episodes

Browse all episodes →