How do I calculate the cost of an API request?

Identify your model, estimate input/output tokens (plus reasoning and image tokens if used), choose Standard or Batch pricing, consider cached input, then plug the numbers into the calculator to get the total.

Is Batch pricing cheaper than Standard?

Batch pricing often reduces per-token cost but suits asynchronous or bulk workloads. Use the calculator to compare exact totals for your scenario.

How do image tokens affect cost?

Vision inputs are billed per the model’s image pricing. Include image tokens along with text tokens to get the full cost.

Use System prompt Optional • collapse to keep focus

System

Max 30,000 characters 0 / 30,000

Prompt

Max 50,000 characters 0 / 50,000

Use Images Vision input • 512×512 ≈ 85 tokens/tile

Upload images

Select up to 10 files • JPEG/PNG/WebP • ≤ 5 MB each • ≤ 4096×4096 px

Add by URL

Remote images must allow CORS to read dimensions • JPEG/PNG/WebP • ≤ 4096×4096 px

Total image tokens: 0

(auto-filled into pricing)

Image tokens (input, total)

Use Response Paste assistant output to estimate completion tokens

Response (optional, used to estimate output tokens)

Max 50,000 characters 0 / 50,000

Reasoning tokens ⓘ Reasoning tokens (for reasoning-capable models) are billed at the output rate. Enter manually if applicable.

Model

Requests (N)

Options

Standard

Batch

Use cached input price Uses the discounted cached input rate for repeated prompts.
Only for models that support input caching.

⚠️ Most optimistic estimate — assumes your entire input is reused from cache.

Text mode: counts tokens from your text; images and reasoning are optional.

Result

Press “Calculate cost”.

Prompt Cost Calculator

Short instruction: Pick your model, count input/output (and reasoning/image) tokens, choose Standard or Batch, toggle cached input if relevant — then calculate the total instantly below.