Hermes 3 70B

The 405B's personality at daily-driver speed.

Nous Research · Llama 3.1 70B · 70B parameters · 131K context · Pro $20/mo · Censorship: none

Hermes 3 70B is the pragmatic pick in the Hermes family: the same neutral-alignment training and amorphous, steerable persona as its 405B sibling, at a size that answers fast enough for everyday conversation. Nous built the Hermes line around a simple idea — the model should become what the user needs, not what a policy team imagined.

In practice that makes it the best general-purpose uncensored 70B for people who want one model for everything: writing, advice, roleplay, research questions, brainstorming. It rarely needs prompt gymnastics; you ask, it answers, in whatever voice you've set.

What Hermes 3 70B is best at

• Steerable identity: Adopts whatever persona or register you define — and keeps it.
• Everyday speed: 70B answers fast enough for rapid-fire conversation.
• No refusal reflex: Neutral alignment from the same recipe as the 405B.
• 131K context: Long chats and big documents without amnesia.

Frequently asked questions

What's the difference between Hermes 3 70B and 405B?

Same training philosophy, different scale. The 70B is markedly faster and covers everyday use; the 405B goes deeper on complex analysis and intricate long-form work. Start at 70B, escalate when a task demands it.

Is Hermes 3 70B good for roleplay?

Very — its steerable, low-ego persona makes it excellent at inhabiting characters, and 131K context keeps long stories coherent.

Is Hermes 3 70B uncensored?

Yes. The Hermes series is trained for neutral alignment — it answers adult and controversial questions directly instead of refusing.

Related models: Hermes 4 · Euryale · Dolphin Venice

Start free → · All models · Pricing