Service tiers

Run the same model at a different speed and price. Choose a tier for each request.

Tier	Price change	Use it for
`fast`	+50%	Interactive requests that need the lowest delay.
`default`	—	Normal application traffic.
`flex`	−30%	Asynchronous work where latency is not the primary concern.

Set the tier

from openai import OpenAI
import os

client = OpenAI(
    base_url="https://api.gradiated.com",
    api_key=os.environ["GRADIATED_API_KEY"],
)

response = client.chat.completions.create(
    model="YOUR_MODEL_ID",
    messages=[{"role": "user", "content": "Summarize this document."}],
    extra_body={"service_tier": "flex"},
)

Set service_tier on each request that needs different behavior. If you omit it, Gradiated uses default.

flex uses the normal request and response API. It does not create a batch job.