How long a bulk LLM job takes — and what it costs — with parallel workers.
A bulk job’s runtime is total tokens ÷ combined throughput, so more workers finish faster — until the provider’s TPM/RPM ceiling bites. This tool shows time and cost together. Check the ceiling on the rate limit calculator and the discount on the batch API savings calculator.
Free batch processing calculator — item count, tokens each, speed and concurrency to the wall-clock time and total cost of a bulk LLM job.
Total tokens = items × tokens per item. Divide by combined throughput (speed per worker × number of parallel workers) to get wall-clock seconds. Ten workers at 60 tok/s deliver 600 tok/s, so an 8M-token job takes about 3.7 hours — before rate limits throttle you.
If the job is not time-sensitive, yes. Batch endpoints (OpenAI, Anthropic, Gemini) run jobs asynchronously within a 24-hour window at roughly half the per-token price. You trade immediate latency for a large discount — ideal for embeddings backfills, evaluations and offline enrichment.