Hermes 3 405B online, no download.

What Hermes 3 405B is

Nous Research took Meta's Llama 3.1 405B and fine-tuned it. That is Hermes 3 405B: 405 billion weights, open weights, and a house style that follows long instructions closely and holds a persona over many turns. It is a plain instruct model, so it answers directly instead of writing a separate reasoning pass first.

Running it yourself is where most people stop. Even at 4-bit, 405 billion weights are hundreds of gigabytes in memory. That is a server job, not a laptop job, and renting the hardware for one evening costs more than a month of Cloud.

What you get

• Chat in the browser. No install, no driver, no VRAM check. Open the Studio, pick the model, type.
• Tools in Agent and Code mode. Our proxy translates tool definitions into the prompt convention this model was trained on and reads the calls back out, so the coding agent works on it like on any other model in the catalog.
• 15 chat models on the same plan, plus the image, video and audio studios. Hermes 3 405B is one pick in the list, not a separate product.
• A personal API key. One endpoint for your editor, your scripts and any OpenAI-compatible tool.

Call it from your own code

Create a key under Cloud API keys in account settings, point the base URL at https://lu-labs.ai/api/inference/v1 and send the catalog model id. Anything that speaks the OpenAI chat format works, including Aider, LibreChat and a plain curl call. The key spends your plan credits and cannot read or change your account.

curl https://lu-labs.ai/api/inference/v1/chat/completions \
  -H "Authorization: Bearer YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"NousResearch/Hermes-3-Llama-3.1-405B",
       "messages":[{"role":"user","content":"Hello"}]}'

How to start in three steps

1. Open checkout and take the Hosted plan. Card details go to Stripe, and the payment is secured with 3D Secure.
2. The Studio opens on your account. Pick Hermes 3 405B in the model picker.
3. Type your first prompt. For scripts, create an API key in account settings and use the base URL above.

What it costs

The honest limits

Cloud is shared hardware. At busy hours a request can wait in a queue for a few seconds before it starts streaming. Credits are one pool for text, image and video together, so a long agent session and a batch of renders draw from the same budget. The catalog is curated: we host open-weight models we have priced and tested, and you cannot upload your own base model. Account data is hosted in the EU and you can delete it yourself. We never train on your data and we never sell it. Hermes 3 405B has no thinking mode, so if you want step by step reasoning, pick one of the reasoners in the same picker.

Questions