Setting up Nemotron 3.5 Super VL in Hermes Agent gives your agent a free model that can actually see: screenshots, PDFs, charts and receipts. It even scored 63.44% on Hermes Bench, Nvidia’s test built for Hermes, beating its own bigger Ultra model. In the video below I run it in my Agent OS. Here’s how to add it to Hermes and what to use it for.
Key takeaways
- Add it to Hermes through OpenRouter (
nvidia/nemotron-3.5-super-vl, with a:freevariant). - Or run it locally with Ollama or LM Studio and point Hermes at it.
- Best for jobs that start with a screenshot, PDF or chart.
- Keep it in its own Hermes profile so you can switch in one click.
Why use Nemotron 3.5 Super VL in Hermes
Hermes is only as good as the model behind it, and most free models are text-only. Nemotron 3.5 Super VL reads images and documents too, supports tool calling, and has up to a 1M-token context — a strong fit for agents. On Hermes Bench it scored 63.44% (Nvidia-reported) versus 56.33% for Nemotron 3 Ultra.
How to set up Nemotron 3.5 Super VL in Hermes Agent
- Create a new Hermes profile and name it after the model — I keep one profile per model.
- Choose OpenRouter as the provider in
hermes model(add your OpenRouter key if you haven’t). - Pick the model
nvidia/nemotron-3.5-super-vl:freefor the free version, ornvidia/nemotron-3.5-super-vlfor the paid one. - Test it — send it a screenshot or PDF and ask a question about it.
Running it locally instead? Start it in Ollama or LM Studio, then add the local server as a custom OpenAI-compatible endpoint in Hermes.
Want this working in your business, not just bookmarked? Using free open models in your AI agents is exactly what we build together inside the AI Profit Boardroom — 3,400+ members, four live calls a week, and a full one-hour DeepSeek Harness course in the classroom.
Prefer it mapped 1-on-1 first? Book a free strategy session and we’ll plan it for your exact situation.
Join the AI Profit Boardroom →Claim Your Free SEO Strategy Session →
What to use it for in Hermes
- Reading client reports and pulling out the numbers.
- Understanding dashboard screenshots and flagging what changed.
- Processing receipts and invoices into structured data.
- Browser tasks where the agent needs to see the page.
- Research with charts — reading graphs in reports and papers.
Nemotron 3.5 Super VL in my Agent OS
In my agentic operating system, Hermes, OpenClaw, OpenCode and Claude all sit side by side, and I can slot a model like this in for the jobs that start with a picture or a file. If you run out of free tokens, switch the profile to another free model and carry on. My best Hermes model guide covers which model to use for what.
Limits to know
Let me be clear: this is not a frontier model. It’s not going to replace Claude Opus 5.5 for your hardest work. What it is: a free, open, fast model that can read screenshots, PDFs and charts and then act on them — which is exactly the job a lot of agent work starts with.
Free variants can be rate limited, so keep a backup model ready.
A visual workflow to try first
Here’s a simple workflow that shows off the vision side. Every Monday, take a screenshot of your analytics dashboard and drop it into a Hermes chat on the Nemotron profile. Ask Hermes to read the numbers, compare them with last week (Hermes remembers), flag anything unusual, and send you a three-line summary on Telegram. Then turn it into a skill so it’s one command next week. It’s exactly the kind of job text-only models can’t do.
Tips for using it well in Hermes
Keep images clear and cropped to what matters — a tight screenshot of a table beats a full-screen capture. Ask specific questions (“what’s the total in the last column?”) rather than “what do you see?”. Use thinking mode only for harder reasoning, since it uses more tokens. And if the free variant is rate limited, switch the profile to the cheap paid version or another free model so your workflow never stops.
Combining it with other models
The smartest Hermes setups route by job. Use Nemotron 3.5 Super VL for anything visual or long-document, a fast free text model for quick research, and a frontier model like Claude Opus 5.5 for the hardest reasoning. Hermes makes switching a one-click job, so each task gets the right brain.
Final thoughts
One more reason to try it: Hermes Bench was designed around how Hermes Agent actually works — tools, multi-step tasks and recovery from mistakes — so a strong score there is more relevant to Hermes users than a general chat benchmark. Nvidia reported 63.44% for Super VL, ahead of its own flagship.
How I’d test it this week
Spend twenty minutes on it this week. Set up the profile, send it three real images from your work — a report, a receipt, a screenshot — and ask it to extract something useful from each. If the answers are right, schedule one of those jobs. You’ll have a free visual assistant running inside Hermes by the end of the day.
If it struggles with a particular kind of image, try a clearer, tighter crop or ask a more specific question before giving up on it.
Troubleshooting Nemotron in Hermes
If Hermes says the model isn’t found, check the exact model id, including the :free suffix for the free version. If replies stop with rate-limit errors, switch to the paid variant or another free model. If images aren’t understood, make sure you’re sending them as images rather than pasting a file path, and that the provider supports vision for that model. And if you’re running locally, check the server’s context size is at least 64K — Hermes needs room for its tools and memory.
The bottom line on Nemotron 3.5 Super VL in Hermes Agent
Adding Nemotron 3.5 Super VL to Hermes Agent takes a few minutes and gives you a free model that can see. Set it up in its own profile and give it your visual jobs.
FAQ: nemotron 3.5 super vl hermes agent
Can I use Nemotron 3.5 Super VL in Hermes Agent?
Yes — via OpenRouter (including the free variant) or a local server from Ollama or LM Studio.
What is the OpenRouter model id?
nvidia/nemotron-3.5-super-vl, or nvidia/nemotron-3.5-super-vl:free for the free version.
How did it score on Hermes Bench?
63.44%, versus 56.33% for Nemotron 3 Ultra (Nvidia-reported).
Can Hermes read images with this model?
Yes — it accepts images, document pages and short videos.
Is it free in Hermes?
The OpenRouter free variant costs nothing, within its rate limits.
Two ways I can help from here. Join the AI Profit Boardroom — it’s where your AI agent model setup gets built with 3,400+ members doing the same.
Or grab a free strategy session and bring your questions.
Join the AI Profit Boardroom →Claim Your Free SEO Strategy Session →
About Julian Goldie
I’m Julian Goldie — founder of Goldie Agency, best-selling author, and an AI educator with 428K+ YouTube subscribers. I’ve spent 10+ years in SEO and online business and I test every AI agent on my own work before I recommend it. Join the AI Profit Boardroom for the daily builds, or book a free strategy session to talk through yours.
Related reading
Last updated October 2026. This page is a living guide to nemotron 3.5 super vl hermes agent — the facts here move fast and it is updated as they do.
