Yes, Nemotron 3.5 Super VL OpenRouter access is live: the model ID is nvidia/nemotron-3.5-super-vl, and there is a free variant at nvidia/nemotron-3.5-super-vl:free. Grab one OpenRouter API key, paste that model ID into any app that supports OpenRouter, and you are running NVIDIA's new vision model without downloading 80GB of weights. This page shows you how it works, where to plug it in, and how to send it images.
📺 Watch: NVIDIA Nemotron 3.5 Super VL Is INSANE! (FREE AI Model)
🔥 Get the Agent OS as a free bonus: AI Profit Boardroom members get the full Agent OS zip, prompt libraries, daily tutorials and weekly live coaching calls. → Get inside
For most people, OpenRouter is the smartest way to use this model. Running it locally needs a big-memory machine. OpenRouter needs a browser and a key. I use it this way inside my own stack, and it is one of the reasons the model is so easy to slot in alongside everything else I run.
If you want my full Agent OS with Hermes, OpenCode, OpenClaw and Claude already wired up and ready for models like this, it is inside the AI Profit Boardroom ($69/mo).
The two Nemotron 3.5 Super VL OpenRouter model IDs
There are two IDs you need to know:
| Model ID | What it is | Best for |
|---|---|---|
| nvidia/nemotron-3.5-super-vl | The standard paid endpoint | Daily work, automations, anything that runs a lot |
| nvidia/nemotron-3.5-super-vl:free | The free variant | Testing, learning, light personal use |
I am not going to quote prices here, because they change. Check the model page on OpenRouter for current pricing, free-tier rate limits and which providers are serving it. The general rule with free tiers: they come with tighter limits, so they are great for testing and not ideal for a job that runs all day.
If you want the full picture on what the model is, read my Nemotron 3.5 Super VL overview. Short version: it is NVIDIA's open multimodal model for agents, 120B total parameters in a mixture-of-experts design with 12B active per token. It reads text, images, documents and video frames, and it writes text back.
How OpenRouter works, in plain English
OpenRouter is a middleman for AI models. You make one account, get one API key, and that key works across hundreds of models from different companies. You do not need a separate account with NVIDIA, a separate one with another lab, and so on.
Here is why that matters:
- One key, many apps. The same key works in Hermes, OpenCode, OpenClaw and most other agent tools.
- Swap models by changing one line. Want to try a different model? Change the model ID. Nothing else.
- OpenAI-compatible. OpenRouter speaks the same API format as OpenAI. So any tool that lets you set a custom base URL and model name can use it.
To get started, go to openrouter.ai, sign up, open the keys section and create an API key. Keep it private. Treat it like a password.
📺 Watch: OpenClaw + Nvidia Nemotron 3 Super + Ollama is INSANE!
Plugging it into your apps
Hermes Agent
Hermes uses OpenRouter by default, so this is the easiest hookup. Run hermes model, pick OpenRouter, paste your key, and choose nvidia/nemotron-3.5-super-vl or the :free version. Inside a running session, type /model to switch. I go deeper on profiles and fallbacks in my Nemotron 3.5 Super VL Hermes Agent setup guide.
OpenCode
OpenCode already lists "Nemotron 3.5 Super VL" in its model drop-down for free. If you want to route it through your own OpenRouter account instead, add OpenRouter as a provider in OpenCode with your key and pick the model ID.
OpenClaw
NVIDIA lists OpenClaw as one of the harnesses that supports the model. Add your OpenRouter key as the provider, then set the model to nvidia/nemotron-3.5-super-vl.
Any OpenAI-compatible tool
For everything else, you need three things:
- Base URL: https://openrouter.ai/api/v1
- API key: your OpenRouter key
- Model name: nvidia/nemotron-3.5-super-vl or nvidia/nemotron-3.5-super-vl:free
Put those three into the tool's settings and you are done. That works for chat apps, no-code automation tools and most agent frameworks.
Not sure which tool to set up first? Do not hand-edit config files. Just ask your AI agent to do it. Paste something like this:
Add OpenRouter as a provider using my existing OpenRouter key, set the model to nvidia/nemotron-3.5-super-vl:free, and send a one-line test message so I know it works. Do not change my other models.
That is how I want people doing this. Let the agent handle the setup. My Agent OS guide shows how the whole workspace fits together.
📺 Watch: New FREE Nvidia Nemotron 3 Super + Ollama
Sending it images
This is the whole point of the VL in the name. VL means vision-language, so it can see.
On OpenRouter, you send images the same way you would with any vision model: as part of the message, either as an image link or an uploaded image. In the OpenRouter chat playground, you can attach an image right in the chat box and ask a question about it. In agent apps like Hermes, you can paste a screenshot into the session in supported terminals.
Good jobs to give it:
- Screenshots. Ask what is wrong with a page, or pull data out of a dashboard.
- Charts. Per NVIDIA, it scores 80.8% on CharXiv chart reasoning, and 84% with Python code interpretation.
- Documents and scans. NVIDIA reports it is among the top open models on OCRBench v2. Good for invoices, forms and PDFs.
- Agent tasks. Per NVIDIA, it scores 63.44% on Hermes Bench and 87.05% on LangChain evals.
Every one of those numbers is NVIDIA's own, and there is no independent replication yet. Run it on your own real tasks before you trust it with anything important. You can see how I test models on real work in Goldie Bench.
If you want help deciding which jobs in your business to route to which model, book a free AI strategy session.
Free vs paid: which one to use
Start with the :free variant. Test it on five real jobs. See if the output is good enough for what you need.
If it is, and you plan to run it a lot, move to the paid ID. Paid endpoints usually give you more headroom, which matters for automations and agents that run on a schedule. Check the OpenRouter page for the current numbers before you commit.
If the free tier runs out mid-task, you have options. Switch to another model on the same key, set a fallback model in your agent app, or come back later. That is the beauty of OpenRouter: one key, so switching is painless. For the full rundown on every free route, see my Nemotron 3.5 Super VL free guide.
Honest take
This is not a frontier model. It is not going to replace Opus 5.5 for your hardest work. But it is free to try, open, fast, and it can read pictures and documents. Through OpenRouter, it is also the easiest way to use it, because you skip the 80-90GB local download completely.
My stack runs Hermes, OpenClaw, OpenCode and Claude side by side, and OpenRouter is what lets me swap new open models in with a one-line change. If you want that exact setup, it comes inside the AI Profit Boardroom: the Agent OS zip, a video tutorial, a 30-day roadmap, updates as new models land, four weekly coaching calls and daily tutorials. Over 3,400 members are already inside.
FAQ
What is the Nemotron 3.5 Super VL model ID on OpenRouter?
It is nvidia/nemotron-3.5-super-vl for the standard endpoint and nvidia/nemotron-3.5-super-vl:free for the free variant.
Is Nemotron 3.5 Super VL free on OpenRouter?
Yes, there is a :free variant. Free tiers come with tighter rate limits, so check the OpenRouter model page for current limits and paid pricing.
Can I send images to it through OpenRouter?
Yes. It is a vision-language model, so you can attach screenshots, charts and documents in the OpenRouter chat or send them through any app that supports vision input.
Which apps can use it through OpenRouter?
Hermes, OpenCode, OpenClaw and any OpenAI-compatible tool that lets you set a base URL, API key and model name. If you want help picking the right setup, grab a free AI strategy session.
That is the full playbook for Nemotron 3.5 Super VL OpenRouter use: one key, two model IDs, plug it into your apps and start with the jobs that begin with a picture.
