Which LLM Jobs Should Run on the Batch API?

OpenAI, Anthropic and Google all price batch work at half the standard rate. The only question is which of your jobs can wait.

Quick answer

Move anything no user is waiting on to batch: evaluations, backfills, enrichment, nightly summaries and scheduled reports. It's half price on all three major providers.

The discount

Batch prices are half the standard prices on OpenAI's, Anthropic's and Google's pricing pages. On Anthropic, the batch discount stacks with caching.

Good candidates

Running evaluation sets.
Backfilling summaries, tags or embeddings for old data.
Enriching CRM or catalog records.
Nightly and weekly reports.
Moderation or classification queues that tolerate a delay.

Bad candidates

Anything a user is watching.
Multi-step agent loops that need each answer before the next call.
Work with a deadline shorter than the batch window.

How to find them

Look for requests from cron jobs, queues and scripts in your logs, and for traffic spikes at fixed times. Then estimate the saving with the batch savings calculator.

Frequently asked questions

How much does the Batch API save?

Half of the standard price on OpenAI, Anthropic and Google, according to their pricing pages.

How long do batch jobs take?

Results come back asynchronously within the provider's batch window, so use batch only for work that can wait.

Sources: OpenAI's API pricing page, Anthropic's pricing page, Google's Gemini API pricing page. Checked October 30, 2026.

Want us to find your 20%?

Book a free 15-minute API cost teardown. Bring last month's provider usage and we'll show you your biggest waste and the first fix we'd ship.

Get a Free 15-Minute API Cost Teardown