Model Pricing

Overview

Aurora tracks token usage and uses model pricing metadata to calculate cost. That cost feeds dashboard analytics and cost reports. In Enterprise profiles, pricing data also drives budget enforcement.

Pricing can come from the built-in model registry, provider metadata, or operator overrides. Overrides are useful for local models, private provider contracts, regional pricing, or models that are not in the public registry.

Pricing fields

Pricing records can include currency, input_per_mtok, output_per_mtok, cached input pricing, cache write pricing, reasoning output pricing, batch pricing, and modality-specific prices for images, audio, characters, requests, or pages.

Set only the fields that apply to the model. Aurora keeps base pricing and operator overrides separate, then shows the effective merged pricing.

Manual management

Use the dashboard model pricing surfaces to inspect effective pricing, add overrides, export overrides, import YAML overrides, and restore pricing backups.

For static local model metadata, define pricing in provider model metadata:

providers:
  local-vllm:
    type: vllm
    base_url: "http://localhost:8000/v1"
    models:
      - id: my-local-model
        metadata:
          pricing:
            currency: USD
            input_per_mtok: 0
            output_per_mtok: 0
 

Server API automation

For endpoint reference see the Admin API section.

curl -X PUT "http://your-aurora-host/admin/api/v1/model-pricing/openai-primary%2Fmy-model" \
  -H "Authorization: Bearer $AURORA_MASTER_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "pricing": {
      "currency": "USD",
      "input_per_mtok": 0.25,
      "output_per_mtok": 1.00
    }
  }'
 

Pricing selectors should match entries from GET /admin/api/v1/model-pricing or provider/model selectors from GET /admin/api/v1/models.

Once pricing is set

This page covers how to set pricing. The next questions — how Aurora turns tokens into dollars, the two-layer cost model, provenance, the recalculation workflow, and reconciling against provider invoices — are documented separately. See Pricing & Cost Calculation.

Related

• Pricing & Cost Calculation — the two-layer model, provenance, recalc, and reconciliation
• Observability — how usage and cost data surface in the dashboard
• Admin API: Model Pricing — every pricing endpoint
• Cache — how the response cache also tracks savings