How Phaseo calculates model pricing

Phaseo tracks the public execution price for a model-provider combination, including the meter type, unit size, currency, and plan context used by the upstream provider or gateway configuration.

We separate database pricing visibility from gateway top-up fees. Public model pricing focuses on what a request costs to execute, while the pricing page explains any platform-level credit purchase fees.

Pricing values are sourced from structured provider pricing rules, provider-model capability mappings, and model metadata stored in the catalog.

When a provider exposes multiple meters for a model, Phaseo keeps the capability-level pricing so the visible page can distinguish text generation, embeddings, image generation, or other billable surfaces where applicable.

Provider price sheets arrive in different units such as per token, per 1M tokens, per image, per second, or per request. Phaseo stores the raw unit and unit size, then calculates consistent display values for the catalog and calculator.

Currency is preserved as published. Where a provider exposes separate input and output rates, those are shown separately rather than collapsed into a blended estimate.

Pricing rules can carry effective start and end dates. When multiple rules exist for the same capability, Phaseo prefers the active rule with the strongest match and the highest operational priority.

When a provider updates a model price, the older rule is retained only as historical context where the data model supports it; the public catalog is meant to reflect the current active price.

Some providers publish incomplete or delayed public pricing for preview models, region-specific SKUs, or enterprise-only offerings. In those cases Phaseo may show partial pricing or omit a comparison until the rule can be verified.

Displayed prices are an operational reference, not a billing contract. Always confirm production billing terms with the provider or your Phaseo commercial plan where necessary.