Catalog

Look up models, context windows, and pass-through prices before you send a request.

ModelCatalog and ModelSpec describe every routed model. estimate_cost multiplies token counts by the spec's prompt and completion rates.

from enroute import ModelCatalog, estimate_cost
from enroute.types import Usage

catalog = ModelCatalog()
spec = catalog.get("openai/gpt-5.6-luna")
print(spec.provider, spec.context_length, spec.pricing)

cost = estimate_cost(Usage.from_counts(1_000, 200), spec)
print(cost)

The hosted catalog is also on this site at /models and as GET /v1/models. Last Labs does not add a routing markup — the number you see is the upstream list price.