Hyperbolic
Hyperbolic
Frontier

Hermes-3-70B (Hyp)

Fine-tuned

Nous Research's Hermes 3, instruction-tuned Llama 3.1 70B with a more open, less filtered tone. Hyperbolic is the default hosted endpoint.

Hermes-3-70B (Hyp) is a frontier AI model from Hyperbolic. It costs $0.400 per million input tokens and $0.400 per million output tokens (blended $0.400/M), with a 128,000-token context window.

INPUT
$0.400/M
per million input tokens
OUTPUT
$0.400/M
per million output tokens
BLENDED 70/30
$0.400/M
unchanged since 3 May
CONTEXT
128,000
tokens
What it is good at
  • Open-weights
  • Less filtered tone than vanilla Llama
  • Strong tool use
Typical use cases
  • Self-hosted chat with neutral tone
  • Roleplay-friendly assistants

Benchmarks

vs. best public score
Scores inherited from Llama 3.1 70B (Hermes 3 base) — this is a hosted variant of the same open-weights model, so the underlying benchmark scores are identical.
MMLU83%
Multitask academic knowledge across 57 subjects.
Graduate-level science questions, "Google-proof".
MATH68%
High-school competition math problems.
Python function synthesis from docstrings.
LMArena Elo1248 Elo
Crowd-sourced head-to-head preference Elo rating.
Hand-curated from each provider's published reports and public leaderboards. Methodology varies across sources, treat as directional rather than authoritative.

More from Hyperbolic

See all 4 →

Frequently asked questions

How much does Hermes-3-70B (Hyp) cost?

Hermes-3-70B (Hyp) costs $0.400 per million input tokens and $0.400 per million output tokens, for a blended reference rate of $0.400 per million tokens.

What is Hermes-3-70B (Hyp)'s context window?

Hermes-3-70B (Hyp) supports up to 128,000 tokens of context in a single request.

What is Hermes-3-70B (Hyp) best for?

Hermes-3-70B (Hyp) is well suited to Open-weights, Less filtered tone than vanilla Llama and Strong tool use.

Who makes Hermes-3-70B (Hyp)?

Hermes-3-70B (Hyp) is developed and served by Hyperbolic. It was released in Aug 2024.

Terms used on this page