OpenRouter's billing protection removes charges for specified failed responses, even when the underlying model provider charges for processing the prompt. The policy covers a response with no generated text tokenThe chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.Full definition → and a blank or null completion status, or a response whose completion status reports an error.
The protection applies automatically across its model catalog. For applications making repeated AI requests, it shifts the provider's processing charge for those failures away from the customer. Protected requests appear with zero cost in the activity record.
Terms in this piece · Glossary
token — The chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.
