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.