Choose an image model when you create a Summon or add an image. The best option depends on whether you value speed, prompt accuracy, graphic-design text, or final-image quality.
Choose a current image model
- Open the Create flow or Add Image screen. Enter the image description and choose an art style.
- Open Image Model. The current selectable models are Flux 2 Flash, Flux 2 Standard, Ideogram V3 Turbo, Ideogram V3 Balanced, and Ideogram V3 Quality.
- Compare the options. Use Flash for quick concepts and rerolls, Standard for stronger general prompt accuracy, and Ideogram for typography, labels, symbols, logos, banners, or graphic-design-style images.
- Review the Mana total. The app calculates the current cost from the model, quality, and number of images before generation.
- Generate the image. Review the result and adjust the description, art style, or Style Chips before rerolling.
Compare Ideogram quality levels
- Turbo prioritizes speed and lower usage.
- Balanced spends more processing on the result.
- Quality provides the highest Ideogram quality level and requires PRO.
Exact Mana values are intentionally not duplicated here. The confirmation total in the app is the source of truth when pricing changes.
Use Style Chips and enhancements
- Write the main image description first. Choose an art style as the broad direction.
- Add optional Style Chips. Use a small number of modifiers that reinforce the subject and mood.
- Choose an enhancement when available. Options depend on the Summon type and current form.
- Review the live Mana total. Optional choices can change the displayed total.
- Generate and review. Remove conflicting chips or simplify the description before trying again.
Type-specific enhancements can include Legendary for Characters, 3D Panorama for Locations, or Magical for eligible Equipment.
Troubleshooting
A model is missing
A model may be hidden during provider maintenance. Choose another visible model and check Community Updates for temporary notices.
The result does not follow the prompt
Simplify competing instructions, choose a model suited to the task, and use Style Chips for concise visual modifiers.