[management] Return default rates with each discovered model

The endpoint reported pricing_known and then made the operator find the
price themselves. The dashboard has nothing to prefill a model row with, so
a discovered model either arrived at zero — silently metering every request
against it as free — or had to be priced by hand against a table NetBird
already ships.

Each model now carries the rates it would actually be billed at.

Rates come from the live default pricing table rather than the compiled-in
catalog, because that is the table the synthesiser ships to the proxy: an
operator running a defaults_llm_pricing.yaml would otherwise be shown one
price in the form and charged another. It is the same lookup the catalog
endpoint prefills from, so a model reached by either route prices
identically — pinned by TestDiscoveredRatesMatchTheCatalogEndpoint, since
the two are separate call paths that would otherwise drift.

pricing_known now derives from that same lookup instead of a second pass
over the compiled catalog, so "we can price this" and "here is the price"
can no longer disagree.

input_per_1k and output_per_1k are required and sent even at zero: an
unpriced model is offered at zero and flagged rather than withheld — the
vendor says the credential can reach it, and hiding it would hide a model
the operator genuinely has. The cache rates stay absent when unset, matching
the catalog response, because a zero there reads as "free" rather than "not
applicable".

shared/management/http/api/openapi.yml

@@ -5381,11 +5381,38 @@ components:
           example: "EU Anthropic Claude Haiku 4.5"
         pricing_known:
           type: boolean
-          description: Whether NetBird's shipped pricing table can price this model. When false the operator must set input/output rates, or requests to it would record a cost of zero.
+          description: Whether NetBird's shipped pricing table can price this model. When false the rates below are all zero and the operator must set them, or requests to this model would record a cost of zero.
           example: true
+        input_per_1k:
+          type: number
+          format: double
+          description: Default input token price per 1k tokens, in USD, from the same table the proxy bills with. Zero when pricing_known is false.
+          example: 0.005
+        output_per_1k:
+          type: number
+          format: double
+          description: Default output token price per 1k tokens, in USD. Zero when pricing_known is false.
+          example: 0.015
+        cached_input_per_1k:
+          type: number
+          format: double
+          description: OpenAI-shape cache rate — default cost per 1k cached prompt tokens (a subset of input tokens), in USD. Absent when the model has no cached-input discount.
+          example: 0.000075
+        cache_read_per_1k:
+          type: number
+          format: double
+          description: Anthropic-shape cache rate — default cost per 1k cache-read tokens (additive to input tokens), in USD. Absent when the model has no cache-read rate.
+          example: 0.0003
+        cache_creation_per_1k:
+          type: number
+          format: double
+          description: Anthropic-shape cache rate — default cost per 1k cache-creation tokens (additive to input tokens), in USD. Absent when the model has no cache-creation rate.
+          example: 0.00375
       required:
         - id
         - pricing_known
+        - input_per_1k
+        - output_per_1k
     AgentNetworkCatalogProvider:
       type: object
       properties:

shared/management/http/api/types.gen.go

@@ -2122,13 +2122,28 @@ type AgentNetworkConsumptionDimensionKind string
 
 // AgentNetworkDiscoveredModel defines model for AgentNetworkDiscoveredModel.
 type AgentNetworkDiscoveredModel struct {
+	// CacheCreationPer1k Anthropic-shape cache rate — default cost per 1k cache-creation tokens (additive to input tokens), in USD. Absent when the model has no cache-creation rate.
+	CacheCreationPer1k *float64 `json:"cache_creation_per_1k,omitempty"`
+
+	// CacheReadPer1k Anthropic-shape cache rate — default cost per 1k cache-read tokens (additive to input tokens), in USD. Absent when the model has no cache-read rate.
+	CacheReadPer1k *float64 `json:"cache_read_per_1k,omitempty"`
+
+	// CachedInputPer1k OpenAI-shape cache rate — default cost per 1k cached prompt tokens (a subset of input tokens), in USD. Absent when the model has no cached-input discount.
+	CachedInputPer1k *float64 `json:"cached_input_per_1k,omitempty"`
+
 	// Id Identifier to register on the provider record, in the form the vendor issues it. For Bedrock this is the region-prefixed inference-profile id, which is the only form AWS accepts at invoke time.
 	Id string `json:"id"`
 
+	// InputPer1k Default input token price per 1k tokens, in USD, from the same table the proxy bills with. Zero when pricing_known is false.
+	InputPer1k float64 `json:"input_per_1k"`
+
 	// Label Vendor-supplied display name, where the vendor supplies one.
 	Label *string `json:"label,omitempty"`
 
-	// PricingKnown Whether NetBird's shipped pricing table can price this model. When false the operator must set input/output rates, or requests to it would record a cost of zero.
+	// OutputPer1k Default output token price per 1k tokens, in USD. Zero when pricing_known is false.
+	OutputPer1k float64 `json:"output_per_1k"`
+
+	// PricingKnown Whether NetBird's shipped pricing table can price this model. When false the rates below are all zero and the operator must set them, or requests to this model would record a cost of zero.
 	PricingKnown bool `json:"pricing_known"`
 }