Service Tiers

Google's Gemini models support selectable processing tiers that trade latency and availability against price. You pick one per request with the OpenAI-compatible service_tier parameter, and LLM Gateway routes it to the right place for the underlying provider.

Using the service_tier parameter

curl -X POST "https://api.llmgateway.io/v1/chat/completions" \
  -H "Authorization: Bearer $LLM_GATEWAY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "google-vertex/gemini-2.5-pro",
    "service_tier": "priority",
    "messages": [
      { "role": "user", "content": "Summarize this incident report." }
    ]
  }'

Accepted values are flex, priority, and default/auto (standard). Providers that don't support tiers ignore the parameter.

Supported providers

Service tiers apply to Google's Gemini models. Under the hood the two Google integrations expose the feature differently, but the service_tier parameter is the same for both:

• Google Vertex AI (google-vertex) — sent as the X-Vertex-AI-LLM-Shared-Request-Type request header. Flex and Priority are served only on the global endpoint, which is the gateway default.
• Gemini Developer API (google-ai-studio) — sent as a service_tier field in the request body.

Billing follows the served tier

Because Google can downgrade a request, LLM Gateway bills the tier that was actually served, not the one you requested:

• A priority request that runs as priority is billed at +80%.
• A flex request that runs as flex is billed at −50%.
• A request that Google downgrades to standard is billed at the standard rate — you are never charged the premium for a request that didn't run as priority.

The served tier is read back from the provider response — Vertex reports it in usageMetadata.trafficType (ON_DEMAND_PRIORITY / ON_DEMAND_FLEX / ON_DEMAND), and the Gemini Developer API reports it in the x-gemini-service-tier response header. The multiplier scales per-token costs (input, output, cached, and image tokens); flat per-request and web-search fees are unaffected.

You can see per-tier pricing for each model on its model page under Processing Tiers.