OpenAI Vision Cost Calculator

Low Detail Images: Fixed cost of 85 tokens regardless of size.

High Detail Images: Cost depends on size.

1. Scale to fit in a 2048px x 2048px square, maintaining aspect ratio
2. Scale so that the image's shortest side is 768px long
3. Count the number of 512px squares needed to cover the image
4. Each 512px square costs 170 tokens
5. Add 85 tokens to the total

Examples:

• 1024 x 1024 image (high detail): 765 tokens
• 2048 x 4096 image (high detail): 1105 tokens
• Any size image (low detail): 85 tokens

All GPT-4.1 models: Cost based on 32px x 32px patches.

1. Calculate 32px x 32px patches needed to cover the image
2. Cap at 1536 patches maximum (scaling down if needed)
3. For GPT-4.1: 1 token per patch
4. For GPT-4.1-mini: 1.62 tokens per patch
5. For GPT-4.1-nano: 2.46 tokens per patch

Examples:

• 1024 x 1024 image: 1024 tokens (GPT-4.1)
• 1800 x 2400 image: 1452 tokens (GPT-4.1)
• Large images are scaled down to stay under 1536 token cap