Playing AI Dungeon
Getting Started
AI Dungeon 101
Adventure AwaitsCommon Questions
About the AI
Account Questions
Memberships & Benefits
Privacy & Security
Terms of Service
Other Questions
Community 
Feedback & Bugs
Product Updates
Table of ContentsAbout our Ultra AI Models
AI Dungeon relies on AI language models to drive game experiences. We also offer several AI image models to generate pictures during gameplay. Each of these models has been developed independently and has different characteristics.
In AI Dungeon, Ultra models are available for Legend and Mythic tiers. These models currently include GPT-4 Turbo, Dalle-3, and Dalle-3 HD.
Ultra models are phenomenal. While they are unquestionably better and more powerful than Premium models, that performance level is only suited for more discerning players who value quality above all else and are willing to pay for the ability to use them.
GPT-4 Turbo
GPT-4 Turbo is the most capable model available in AI Dungeon. It was specifically designed with improved comprehension and higher-quality generations in mind. With up to 128k tokens of context, GPT-4 Turbo could recall almost all of your Adventure, Memory, Story Cards, and more. This amount of context is absolutely absurd—that’s about 160-200 pages of text, and only our most dedicated players have Adventures of that length.
Legend and Mythic players can spend 1 Credit per action for every 250 tokens of GPT-4 Turbo context up to 128k tokens. Mythic includes unlimited actions on GPT-4 Turbo with 1k tokens as part of the subscription and will only be charged Credits for context used beyond 1k. Filling up the entire 128k context will cost over 500 Credits every turn.
Dalle-3
Dalle-3 is a powerful image model capable of generating highly detailed images in a wide variety of styles. It is highly effective at including particular details mentioned in prompts and reducing the number of nonsensical abnormalities in generated images. Dalle-3 images cost 15 Credits per generation.
Dalle-3 HD
Dalle-3 HD is a high-definition version of Dalle-3, capable of generating images at an even higher resolution with better consistency and accuracy. It has a more sophisticated understanding of complex text prompts, enabling pictures that more closely align with players’ intent and creativity. Dalle-3 HD is also extremely good at generating text in images when prompted. These images cost 30 Credits per generation.
