Unverified50% confidenceFactExact time
Provisioned Throughput is an hourly-billed capacity reservation model in which users purchase fixed input and output token processing capacity for a specific model (measured in Model Units), usually with a minimum commitment period and strongly tied to a specific model version.
1
Sources
50%
Confidence
Long-term
Relevance
10/18/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe prefill phase processes the complete input context, computes attention for all input tokens, and generates the KV cache; it is a compute-intensive operation whose time consumption is strongly correlated with input length.68% similarUnverifiedThe prefill phase requires processing all input tokens at once and is a compute-intensive operation, whereas the decode phase generates only one token per step and is a memory-bandwidth-intensive operation.67% similarUnverifiedLarge model services adopt a Token quota mechanism, where each model call consumes Tokens measured by the input and output text67% similarUnverifiedLarge model service providers typically use batching techniques to improve throughput; continuous batching allows new requests to be dynamically inserted into batches being processed, whereas static batching requires waiting for a complete batch to finish processing.67% similarUnverifiedtoken消耗的精确计量依赖于对Cursor官方API响应体中usage字段(包含prompt_tokens、completion_tokens)的实时解析66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1060483API
curl https://kongchang.com/api/v1/knowledge/claims/1060483MCP
get_claim(id=1060483)