MythoMax 13B

MythoMax 13B is a community-built fine-tune of Meta's Llama 2 13B, released by the creator Gryphe as a model in the 13-billion-parameter class. According to its third-party listing, it is described as one of the highest-performing and most popular fine-tunes of the Llama 2 13B base, with tuning emphasis on rich descriptions and roleplay-style generation. The model operates as a text-to-text system, exchanging input and output purely as text, which makes it well suited to conversational interfaces, narrative writing, and dialogue-driven applications.

The model's practical profile centers on a compact 4K-token context window for both input and output, an efficient tier with fast response speed, and an input cost near the budget end of hosted inference. It is a reasonable fit for users who want Llama-2-class fluency with a more characterful, descriptive voice, particularly for short roleplay scenes, creative writing prompts, and chat-style assistants. Its smaller context budget means it is best matched to focused, single-turn or short multi-turn exchanges rather than long document analysis, and its lack of vision input keeps it firmly in the text-generation niche.