WorldFlow AI Inference Business Case

Workload

Workload template

Requests per day

Input tokens

Output tokens

Work suited for private AI

Repeatable work

Meets private-model quality

Provider and deployment choices

Frontier provider and model

Frontier input price per 1M tokens

Frontier output price per 1M tokens

Open-weight model and hardware

Deployment

GPU hourly cost

Advanced calibration

Scenario sensitivity

Business-case mode

Traffic and pricing

Cached input, %

Cached input price per 1M tokens

Serving and reuse

Input throughput per GPU

Output throughput per GPU

Throughput basis

Max util., %

Minimum GPUs

Peak traffic multiplier

Resilience GPUs

Exact-prefix hit, %

Semantic hit, %

Manual GPU override

Owned infrastructure

Purchase cost per GPU

Useful life, months

Power, kW/GPU

Electricity cost per kWh

Financial

One-time evaluation and migration

Monthly operating cost

Monthly software cost

Scenario name