GPT Image 2 series

Three quality tiers

GPT Image 2 is available in Low, Medium, and High quality tiers. Each tier supports text-to-image generation and image editing.

Use this page to compare the family variants and input modes. Use Costs by model and task when you need the current credit table.

When to choose GPT Image 2

• Use Low when you want faster, lower-cost visual exploration.
• Use Medium as the balanced default for most image generation and edit tasks.
• Use High when final quality matters more than iteration cost.
• Use the edit variants when you need reference-image changes instead of fresh generation.

Related KrafLayer docs

• [Image generation models](/docs/image-generation)
• [Nano Banana series](/docs/nano-banana-models)
• [Costs by model and task](/docs/generation-cost)
• [Plans and price](/docs/pricing-and-credits)