- Filter is the friendly name displayed on Cost Report filters in the Vantage console.
- API Field Name is the fully-qualified field reference used in API requests and VQL queries (for example,
costs.account_idortags.name). Use this exact value when writing VQL. - Data Type indicates the type of data accepted for the field.
- Example Value is a representative value to give you a sense of what the field actually contains. This is illustrative, and your data may use different values.
Usage Unit
Usage Unit is a shared field across cost providers. It identifies the unit used to measure a line item’s consumption, such as hours, storage, requests, or tokens.
The Usage Unit filter is available for individual providers, Custom Providers, and All Providers when the selected provider context contains at least two distinct, non-empty units. If the context contains zero or one unit, Vantage omits the filter because it would not further narrow the report. See Supported Providers for the providers that populate usage data and examples of their units.
Vantage normalizes Usage Unit labels in the console for consistent spelling and capitalization. When querying with VQL, use the exact underlying value shown by View as VQL. See Usage-Based Reporting for more information.
Provider Field Values
Each cost provider also has a normalized name within the API. Theprovider field is part of the costs namespace.
To reference a Custom Provider in VQL queries, navigate to the Integrations page. Use the displayed Provider ID (e.g.,
custom_provider:accss_crdntl_123a45bfdaf38765).AWS
Vantage also applies Managed AI Tags to supported AI provider costs: a normalized
vntg:ai: tag schema (for example, vntg:ai:model, vntg:ai:model_provider, and vntg:ai:token_type) that is available as tags.name values so you can analyze AI spend with consistent dimensions across providers.
When AWS Bedrock LLM Enrichment is enabled, the per-request requestMetadata keys your application attaches are available as tags.name values on enriched Amazon Bedrock cost rows, alongside vntg:ai: tags such as vntg:ai:principal_id, vntg:ai:operation, and vntg:ai:inference_scope.
When Custom LLM Enrichment is enabled, the per-request tags you deliver (plus vntg:ai: tags such as vntg:ai:model) are available as tags.name values on enriched model-provider cost rows.
Anthropic
Anthropic costs come from two data sources—the Claude API (Admin API key) and Claude.ai (Claude Enterprise Analytics API)—that share the same fields but populate them differently. See the Anthropic integration documentation for a full per-source breakdown. Use theanthropic:billing_source tag (admin or analytics) to distinguish the two.
Anyscale
Azure
Azure CSP
Azure CSP reporting dimensions differ slightly from standard Azure integrations.
Baseten
CircleCI
ClickHouse Cloud
Cloudflare
Confluent
Coralogix
Cursor
Custom Providers
Available fields are based on what you’ve uploaded for this Custom Provider. See the Custom Providers documentation for details.
* Regular and virtual tags associated with this provider
Databricks
Datadog
Elastic
ElevenLabs
Fastly
Fireworks AI
GitHub
Google Cloud
Grafana Cloud
Kubernetes
Linode
MongoDB Atlas
Modal
New Relic
OpenAI
OpenAI cost data spans two data sources, the OpenAI Platform and OpenAI Codex, under the
open_ai provider. Use the openai:billing_source tag (admin for OpenAI Platform or codex_analytics for OpenAI Codex) to distinguish the two. See the OpenAI integration documentation for the full list of dimensions and tags per data source.