Method & evidence
EffortFrontier is a digital Software & SaaS product that sells measurement reports. The financial model and sources that justify pricing and ICP live in the open repository that ships with this product.
What we measure
On your prompts: combinations of model, reasoning_effort, and max_tokens. Output: a PDF cost–quality frontier with confidence intervals and config recommendations. We sell measurement because vendors have little incentive to prove that lower effort is enough, and free tools already cover generic prompt optimization.
Digital fulfillment
Purchase grants a report SKU. The PDF is delivered asynchronously (typically within two business days). This is product fulfillment, not a consulting engagement. Support contact: wangqiyuanbistu@gmail.com.
Payback illustration
If a report costs $149 and you expect ~20% monthly bill reduction, a one-month payback implies a monthly LLM bill near $745. At 5% savings the bar is near $2,980; at 40%, near $373. These are arithmetic illustrations from the BP model, not promises about your workload.
Reproduce the BP
In the parent repository: python model/run_all.py writes outputs/results.json. Sources are in data/sources.json. Unverified claims are labeled.