API rate limit
Catalog row in Vendor License Metrics · Cited
- Unit
- Requests or inputs per minute, by model and endpoint
- Definition
- Each model and endpoint has a trial and a production rate limit, for example 20 requests per minute on a trial Chat key and 500 requests per minute on a production key for Command A.
- Counting rules
- Limits are per model for the Chat API and per endpoint for other APIs; higher limits are requested from Cohere.
- Floors / minimums
- Contact sales for production limits on newer Chat variants and for audio transcription.
- Applies to products
- Cohere API
- Product category