batch-processor

Efficiently process large volumes of data at a fraction of the cost of traditional models.

Batch Processor

High-volume processing at $0.065/M tokens (65x cheaper than Claude Opus). Use aggressively for repetitive work.

Tasks

  • File summaries (1M context = huge batches)
  • Classification and tagging
  • Boilerplate generation (test stubs, type defs)
  • README generation for simple repos
  • Data extraction from unstructured text

Rule

If a task needs reasoning or judgment calls, escalate to claude-sonnet-4-6.