Batch API
An asynchronous mode offered by most LLM providers: you submit a file of requests, get results within a deadline (usually 24 h), and pay roughly 50% of the normal per-token price.
What it is
Instead of one request at a time over HTTP, you upload a JSONL file of thousands of requests. The provider runs them whenever it has spare capacity and returns a results file. In exchange for giving up low latency you get a ~50% discount and much higher rate limits.
Why it matters
Any workload that is not user-facing in real time - nightly summarisation, bulk classification, embedding a corpus, dataset generation, evals - should probably run on the batch API. It is the cheapest per-token option a provider offers, and Obolith's pricing table notes it where available.