Input tokens
The tokens sent into the model, including prompts, system instructions, and context.
Long-tail SEO page
AI cloud pricing glossary
Glossary pages support search visibility, clarify buyer confusion, and create dense internal links across the site.
Output tokens
The tokens generated by the model in response to a request.
Cached tokens
Reusable prompt segments that some providers may price differently from fresh input tokens.
Context window
The maximum amount of input and output context a model can handle in one request.
Rate limit
A provider-imposed cap on requests, tokens, or throughput over a given time period.
Latency
The time it takes for a model response to be generated and returned to the application.
Throughput
The amount of model work a system can process over time, often relevant for high-volume apps.
Fallback model
A secondary model used when the primary model fails, is rate-limited, or is too expensive for a given task.