Suscripción vs API

Compare una suscripción de IA con la API de pago por uso según su nivel de uso. Las suscripciones se facturan por ventana de tiempo y no por token, así que no existe un punto de equilibrio exacto: los resultados se dan como rangos, junto con la sensibilidad a la caché y la fecha en que se verificó cada plan.

Los precios y cupos de los planes se mantienen a mano y no se han verificado con cada proveedor por separado. Los precios de API se leen del catálogo, por lo que cambiar la base de precio puede invertir la conclusión. Las reglas de cupo cambian con frecuencia; la página oficial tiene prioridad.

Prompts / día

Tokens de entrada / llamada

Tokens de salida / llamada

Tasa de aciertos de caché 70%

Llamadas al modelo por prompt: 8

En chat es cerca de 1. Un agente planifica, busca, invoca herramientas y resume, así que un prompt suele desencadenar de 5 a 30 llamadas. Por eso un modelo con precio bajo por token puede resultar caro por tarea.

Base del precio de API

Precio de lista de primera parte, la alternativa comparable a una suscripción de primera parte.

Escenarios predefinidos

≈ 4,800 llamadas al modelo / mes con este uso.

Los rangos de coste provienen de la incertidumbre de la caché (±15 pp). Los rangos de capacidad provienen de leer cada cupo de ventana de forma conservadora u optimista. El punto de equilibrio es una región, no un punto, porque los cupos de suscripción son ventanas de tiempo que no se reducen a un número fijo de tokens.

Las ventanas móviles de menos de un día se normalizan suponiendo 176 horas activas al mes (8 h × 22 días); los límites diarios, semanales y mensuales son ventanas de calendario y no se descuentan.

Planes sin cupo calculable

Estos no publican un cupo absoluto, o facturan en una unidad que no puede traducirse a llamadas al modelo. Se enumeran aquí en lugar de asignarles una cifra inventada.

OpenAI FreeGratisModel and feature limits are dynamic and no absolute monthly or message cap is published.
OpenAI Go8,00 US$Higher limits than Free are published without an absolute message or monthly allowance.
OpenAI Plus20,00 US$Advanced-model and feature limits are dynamic; the pricing page gives no absolute API-call equivalent.
OpenAI Pro $100100,00 US$The 5× usage statement is relative to Plus and cannot be converted to an absolute API-call allowance.
OpenAI Pro $200200,00 US$The 20× usage statement has no absolute API-call equivalent; new subscriptions and upgrades are temporarily paused.
OpenAI Business Standard25,00 US$Business usage is subject to service guardrails and has no absolute published API-call allowance.
OpenAI Business Premium125,00 US$Premium usage is relative to the Standard seat and has no absolute published API-call allowance.
Anthropic FreeGratisUsage depends on the model, conversation and feature; no absolute cap is published.
Anthropic Pro20,00 US$Usage credits and weekly limits depend on task and model; no absolute API-call allowance is published.
Anthropic Max 5x100,00 US$The 5× statement is relative to Pro and task-dependent, not an absolute API-call allowance.
Anthropic Max 20x200,00 US$The 20× statement is relative to Pro and task-dependent, not an absolute API-call allowance.
Anthropic Team25,00 US$Per-seat usage is not published as an absolute API-call allowance.
Google AI Plus4,99 US$The plan states a relative 2× usage limit; compute-based limits cannot be converted to API calls.
Google AI Pro19,99 US$The plan states a relative 4× usage limit; compute-based limits cannot be converted to API calls.
Google AI Ultra 5x99,99 US$Usage is measured as compute and varies with the prompt, features and conversation; no API-call cap is published.
Google AI Ultra 20x199,99 US$Usage is measured as compute and varies with the prompt, features and conversation; no API-call cap is published.
xAI FreeGratisRate limits are not published as absolute model-call numbers.
xAI SuperGrok30,00 US$Higher rate limits are not published as absolute model-call numbers.
xAI SuperGrok Plus100,00 US$Higher rate limits are not published as absolute model-call numbers.
Perplexity FreeGratisPro-search and feature limits are not equivalent to a fixed model-call allowance.
Perplexity Pro20,00 US$Search and research usage bundles retrieval, tools and generation, so it cannot map to API model calls.
Perplexity Max200,00 US$Monthly AI credits vary in cost by research feature and model, so they cannot map to API model calls.
Mistral FreeGratisVibe feature limits and Studio credits do not specify a fixed model-call allowance.
Mistral Pro14,99 US$Vibe feature limits and included Studio API credits do not specify a fixed model-call allowance.
Mistral Team24,99 US$Per-seat Vibe use and included credits do not specify a fixed model-call allowance.
GitHub FreeGratisChat, agent and completion limits are product-specific and are not API model-call limits.
GitHub Pro10,00 US$Included GitHub AI Credits have feature- and model-dependent costs, not a fixed API-call conversion.
GitHub Pro+39,00 US$Included GitHub AI Credits have feature- and model-dependent costs, not a fixed API-call conversion.
GitHub Max100,00 US$The $100 monthly AI Credit allocation has feature- and model-dependent costs, not a fixed API-call conversion.
GitHub Business19,00 US$Included GitHub AI Credits have feature- and model-dependent costs, not a fixed API-call conversion.
GitHub Enterprise39,00 US$Included GitHub AI Credits have feature- and model-dependent costs, not a fixed API-call conversion.
Cursor HobbyGratisAgent-request limits are not published as absolute model-call numbers.
Cursor Pro20,00 US$Agent limits and usage-based features do not specify a fixed model-call allowance.
Cursor Pro+60,00 US$Agent limits and usage-based features do not specify a fixed model-call allowance.
Cursor Ultra200,00 US$Agent limits and usage-based features do not specify a fixed model-call allowance.
Cursor Teams Standard40,00 US$Team use and usage-based features do not specify a fixed model-call allowance.