Set Up the OpenAI API for Photo Supreme

OpenAI Pricing

OpenAI isn’t completely free. To get started, you’ll need to make an OpenAI account, and buy some credits for your account. These credits are used whenever Photo Supreme asks OpenAI to analyze an image.

The good news is that OpenAI’s prices are reasonably low. You can check the latest prices here: https://openai.com/api/pricing/

Photo Supreme lets you choose which OpenAI model to use in its preference settings (section Other Settings). Different models have different prices.

The cheapest AI model you can use in Photo Supreme is GPT-5-Nano.

For GPT-5-Nano, OpenAI charges 5 cents ($0.05) for 1 million pieces of text, which OpenAI calls tokens.

When Photo Supreme analyzes an image, it typically uses about 1000 tokens per image.

Let’s calculate how many images you can analyze with 1 million tokens:

• 1,000,000 tokens ÷ 1000 tokens per image = 1,000 images

So, for $0.05, you can analyze about 1,000 images.

Now let’s scale that up to $1:

• $1 is 20 × $0.05
• 1,000 images × 20 = 20,000 images

So, for $1, you can analyze about 20,000 images using GPT-5-Nano.

Cost per image

• $1 ÷ 20,000 images = $0.00005 per image

That’s five ten-thousandths of a dollar, or far less than one-hundredth of a cent per image.

Overview: Number of images you can analyze for $1

While GPT-5-Nano is the least expensive option, GPT-5-Mini is often the best overall choice for most users. It offers a good balance between cost and quality, producing more accurate descriptions, better labels, and more consistent results than Nano, while still being very affordable. For most Photo Supreme users, GPT-5-Mini provides the “sweet spot” where you get noticeably better results without a large increase in cost.