Batch API
Domain: LLM Concepts
π― Motivation
For non-urgent tasks, batch APIs let you submit many requests at once and get results later, at a lower cost.
π Concrete Example
Submit 10,000 product descriptions to summarize. Batch API processes them within 24 hours at half the cost. Streaming would be too expensive and slow for this volume.
π Analogy
Bulk mail vs express delivery - express (streaming) is fast but expensive per item. Bulk mail (batch) is cheap but takes days. For large volumes that arenβt urgent, batch is the smart choice.