Message Batches API
Message Batches API processes up to 10,000 requests per batch asynchronously within 24 hours at a 50% discount on input and output tokens.
An offline-inference path for non-urgent bulk jobs. Half-price tokens in exchange for completion within 24 hours (often faster). Launched for Claude 3.5 Sonnet, Claude 3 Opus and Claude 3 Haiku.
- Date
- Tuesday, 8 October 2024
- Lab
- Anthropic
- Kind
- infra
- Access
- closed API
- Price
- 50% off standard token prices
Figures
| Measure | Value | Measured by |
|---|---|---|
| Batch discount | 50% input and output tokens | company |
| Max requests per batch | 10,000 completion within 24 hours | company |
Later raised to up to 300k output tokens per batch request with a beta header on newer models (2026-03-30 API release notes).
Sources
This record was checked against its sources on 6 October 2026. How we check