Skip to content

status.openkey.ai is live — independent, worldwide uptime proof See it now ->

openkey

Gryphe

1 model built

Gryphe is an independent fine-tuner best known for MythoMax-L2-13b, one of the most-used roleplay models in the open community, along with the Pantheon series. Unlike the large labs on OpenKey, Gryphe isn't a company shipping a broad model family — it's a single creator focused on one thing: tuning Llama 2 architectures for character-driven, long-form roleplay and creative writing. On OpenKey, Gryphe's lineup is a single model, MythoMax 13B, reflecting that narrow, community-driven focus rather than a commercial product roadmap.

Lineup

There's no lineup to navigate here — Gryphe has exactly one model on OpenKey: MythoMax 13B, released July 2023. It runs at $0.06 per 1M input tokens and $0.06 per 1M output tokens (OpenKey price: $0.0618 per 1M, after the flat 3% fee), with a 4,096-token context window. If you want MythoMax-style roleplay output, this is the only option to reach for; there's no flagship-vs-cheap tradeoff to make within Gryphe's catalog.

  • MythoMax 13B ctx 4K · in $0.062 · out $0.062

Prices per 1M tokens, flat 3% fee included.

Questions

What does MythoMax 13B cost on OpenKey?
MythoMax 13B is priced at $0.06 per 1M input tokens and $0.06 per 1M output tokens from the provider. OpenKey adds a flat 3% fee, so you pay $0.06 × 1.03 = $0.0618 per 1M tokens for both input and output. There's no separate cheap or flagship tier — it's Gryphe's only model.
Is MythoMax 13B a free model?
No. Gryphe has zero free models on OpenKey. MythoMax 13B is priced at $0.06 per 1M tokens for both input and output (before OpenKey's 3% fee), so budget for a small per-token cost rather than treating it as a free-tier option.
What is MythoMax 13B good for?
MythoMax-L2-13b is one of the most-used roleplay models in the open community, and it's the model behind Gryphe's reputation, alongside the Pantheon series. Use it for character-driven, conversational, and creative-writing tasks rather than general reasoning or coding — that's not what it was tuned for.