Caching can reuse earlier input processing. Availability, lifetime and billing depend on the endpoint. A cache hit still generates a new response.
01 / INITIAL INPUTProcess & retain
The system processes eligible input and retains reusable state.
02 / LATER REQUESTCheck available state
Reuse depends on matching input, session state and the cache policy.
03 / CACHE HITReuse & generate
Reusing state can reduce repeated work. Any charges follow the billing rules below.
Upstage API
Not verified
No verified caching contract for this exact document model or service. No text-token cache behavior inferred.
How it works
Not collected
No verified caching contract for this exact document model or service. No text-token cache behavior inferred.
Minimum reusable input
Not collected
No verified caching contract for this exact document model or service. No text-token cache behavior inferred.
Lifetime & refresh
Not collected
No verified caching contract for this exact document model or service. No text-token cache behavior inferred.
How to enable it
Not collected
No verified caching contract for this exact document model or service. No text-token cache behavior inferred.
Usage fields
Not collected
No verified caching contract for this exact document model or service. No text-token cache behavior inferred.
Billing rules
Not collected
No verified caching contract for this exact document model or service. No text-token cache behavior inferred.
Limits & data boundaries
Not collected
No verified caching contract for this exact document model or service. No text-token cache behavior inferred.
No verified cache rate for this endpoint. An empty rate is not a zero charge.
Caching reviewed 14 Oct 2026
What the measurements say
No benchmark observations have been added for this model. We do not turn a vendor description into a performance score.
Latency and throughput depend on the host, request size, reasoning settings and load. There is no measured speed or global usage figure in this record.
The record behind the model
Why it is included
Distribution inference: the provider explicitly describes these named prebuilt extractors as document-specific fine-tunes commercially served on top of its OCR. No matching trained weights were identified in the official public inventory. Public OCR/table-related research artifacts do not establish release of these exact commercial fine-tunes.