Hermes 3 405B

Frontier scale. Zero leash.

Nous Research · Llama 3.1 405B · 405B parameters · 131K context · Pro $20/mo · Censorship: none

Hermes 3 405B is the largest genuinely uncensored model you can use without owning a datacenter: Nous Research's full fine-tune of Llama 3.1 405B, trained on their neutral-alignment recipe. It has the depth you expect at frontier scale — nuance, world-knowledge, long coherent arguments — without the personality of a compliance officer.

This is the model for the hard stuff: intricate fiction with dozens of moving characters, dense analysis, ethically complicated questions that smaller models flatten into platitudes. It's slower and heavier than the 70B class, which is exactly the trade you want when the answer matters more than the latency.

What Hermes 3 405B is best at

• 405B parameters: Frontier-class depth — the largest model in the uncensored lineup.
• Neutral alignment: Nous's user-sovereign training, applied at maximum scale.
• Long-form coherence: Holds plots, arguments, and personas across very long sessions.
• 131K context: Entire manuscripts or research dossiers in one conversation.

Frequently asked questions

Is Hermes 3 405B the biggest uncensored model?

Among models you can actually chat with online, effectively yes — it's a full neutral-alignment fine-tune of Llama 3.1 405B, the largest open-weights base of its generation.

When should I use 405B instead of a 70B model?

When depth beats speed: complex analysis, long intricate fiction, or nuanced judgment calls. For quick back-and-forth, Hermes 3 70B or Dolphin Venice feel snappier.

Do I need special hardware to run Hermes 3 405B?

Running it yourself takes hundreds of GB of VRAM. On OpenRogue it runs in the cloud — you just pick it from the dropdown and chat in your browser with a Pro plan.

Related models: Hermes 4 · WizardLM-2 · Magnum v4

Start free → · All models · Pricing