AI

AI Batch Cost Calculator

Plan the cost of bulk AI jobs such as classification, enrichment, summarisation, extraction, and migration runs.

App view

Inputs

PricingChoose a model or enter custom rates
Batch sizeItems, tokens, and retries
Results update automatically

Results

Result Enter values to calculate.
Your result explanation will appear here.

Useful next checks

  • Check the inputs before relying on the result.
  • Try a second scenario to compare outcomes.
  • Read the guide below for context.

# AI Batch Cost Calculator

Use this calculator to estimate the cost of bulk AI processing jobs, such as classification, extraction, summarisation, tagging, enrichment, or data migration.

How It Works

The calculator uses:

  • Billable attempts = source items x (1 + retry rate)
  • Input tokens = billable attempts x input tokens per item
  • Output tokens = billable attempts x output tokens per item
  • Total cost = input token cost + output token cost
  • Cost per source item = total cost / original item count

Retry Rate

Retry rate is a practical buffer for failed requests, malformed responses, quality-control reruns, and duplicate processing. A 5% retry rate means 10,000 source items are estimated as 10,500 billable attempts.

Batch Pricing

Some providers offer separate batch APIs or discounted asynchronous processing. This V1 calculator uses standard text-token rates by default. If you have discounted batch rates, enter them as custom input and output prices.

Limitations

The estimate excludes cache discounts, provider batch discounts unless entered manually, tools, search, file storage, image/audio charges, taxes, regional uplifts, free tiers, and account-specific discounts.

Related Calculators

Frequently asked questions

Does this include provider batch discounts?

No. V1 models standard text-token pricing. If you receive a batch discount, enter the discounted rates as custom input and output prices.

What does retry rate mean?

Retry rate adds extra billable attempts for failed, malformed, or reprocessed items. A 5% retry rate means 10,000 source items are estimated as 10,500 billable attempts.

Should I include prompt overhead?

Yes. Input tokens per item should include the instruction prompt, item data, schema instructions, and any retrieved context sent with each item.