Image Generation API Cost Calculator
Image APIs price per generated image, and the quality tier you pick changes the bill by up to 35×. Enter your monthly volume and retries to see the real per-image and monthly cost across GPT Image models, FLUX, and any custom per-image price.
How image API pricing works
OpenAI's GPT Image models bill per output image with three quality tiers. The tier sets how many tokens the image costs to produce — low quality 1024×1024 is about 272 output tokens, medium about 1,056, and high about 4,160 — so moving from low to high quality multiplies the per-image price by roughly 15× on the legacy models, and by 35× on GPT Image 2 square (its $0.006 low tier is unusually cheap).
FLUX models on hosting providers like fal.ai or Replicate bill a flat price per image with no quality tiers, which makes them dramatically cheaper for volume drafts. Aspect ratio changes the token count, and the effect is not a simple multiplier: OpenAI publishes a separate price table per size, and on GPT Image 2 the portrait and landscape outputs are actually cheaper than square at medium and high quality ($0.041/$0.165 vs $0.053/$0.211), while older models charge more for the larger canvas.
Reference per-image prices (1024×1024)
| Model | Low | Medium | High |
|---|---|---|---|
| GPT Image 2 | $0.006 | $0.053 | $0.211 |
| GPT Image 1.5 | $0.009 | $0.034 | $0.133 |
| GPT Image 1 Mini | $0.005 | $0.011 | $0.036 |
| GPT Image 1 (legacy) | $0.011 | $0.042 | $0.167 |
| FLUX.1 schnell | $0.003 flat | ||
| FLUX.1 dev | $0.025 flat | ||
Prices are list rates for one square image, checked against OpenAI's image generation guide and fal.ai's published GPT Image 2 pricing in September 2026, and they change over time — always confirm against the provider's current pricing page before budgeting.
Methodology: list prices per output image; retries modeled as additional paid generations per final. Byline: KickLLM research. Last reviewed 2026-09-29.
Where the money actually goes
- Retries dominate. A 1.4 attempts-per-final rate adds 40% to every line item; prompt-only iteration on a cheap tier is usually cheaper than re-rolling a high tier.
- Tier discipline saves the most. Draft on Low or Mini, promote to Medium/High only for finals — teams commonly cut image spend by half or more this way.
- Aspect ratios compound. Generating everything in portrait at High quality can cost over 27× a square Low image.