Pricing

Flex pricing

Run the same model at a lower price on spare provider capacity, for work that can wait.

Lower token priceRequests can queueSame model output

What is Flex pricing?

Flex pricing is a lower-priced service tier for AI models. Your request runs on the provider's spare capacity, so the price per token is lower than Standard. Responses can take longer to start because requests may queue. On Spro you turn it on with a model suffix such as :flex, or with the service_tier parameter.

Not every request needs an answer in a second. Classification, document extraction, evaluations and overnight jobs can wait a little longer for a lower bill. Flex runs the same model with the same output quality, and only speed and price change.

Spro offers four service tiers: Standard, the default; Flex, the lowest price; Priority, lower latency at a higher price; and Ultrafast, the lowest latency at the highest price. Each model page lists the tiers it offers, with prices for each.

Key features

What this feature does for your requests.

  • Turn Flex on with one suffix, :flex, or the service_tier parameter
  • Same model and output as Standard, at a lower token price
  • Flex prices shown per model and provider on the model page
  • Usage is billed at the tier that ran, and logged as its own model
  • Fall back from Flex to Standard with a models list

How it works

4 steps, from setup to the first request.

  1. 01

    Pick a Flex model

    Open the model page and check that it has a Flex tab.

  2. 02

    Add :flex

    Add the suffix to the model id, or send service_tier set to flex.

  3. 03

    Allow for queue time

    Set your client timeout to cover the wait before the request starts.

  4. 04

    Compare spend

    Filter the usage log by model to compare Flex and Standard cost.

Is it the right fit?

Specific situations where this feature helps, and where it does not.

Good fit if

  • Nightly pipelines, backfills and evaluation runs that nobody waits on
  • Bulk classification or summarisation where cost matters more than seconds
  • Scheduled background agents

Not a good fit if

  • Chat, voice or any streamed response a person is waiting on
  • Agent loops where each step waits for the previous one
  • Jobs with a hard deadline that a queue could miss

Flex pricing FAQs

Common questions about Flex pricing.

Route every AI request through SPRO.

One OpenAI-compatible API for 350+ models, with intelligent routing, 29 built-in guardrails, observability, and automatic failover.

Talk to sales