Default
General-purpose traffic when you want the normal public offer path.
Models
Choose models by workload, delivery lane, and public price.
Public model ids are the names your agents, workflows, and scripts send in requests. Delivery lanes describe the preference used to serve that model and calculate public pricing.
Model decision pathMove from model id to delivery lane to price dimension using the public catalog.
Find the model idUse the exact public id your agent or workflow will send.
Choose the laneMatch workload intent before comparing rates.
Check the rate tableConfirm text, image, or per-image dimensions before traffic.
Delivery lanes
Pick the lane that matches the workload.
A delivery lane is the public serving preference customers can use when deciding how a workload should balance spend, priority, and availability.
Economy
Cost-sensitive workflows that can trade priority for lower spend.
Coding
Coding agent runs where token-heavy edits need a dedicated delivery lane.
Premium
Latency-sensitive or higher-priority workloads where price is not the first constraint.
Spot
Opportunistic traffic that can accept availability and policy constraints.
Model prices
Published model catalog
Public models are shown from the current public catalog. Use Pricing for the model-rate table before sustained traffic.
Loading dataWaiting for the latest data to become available.
Next step
Turn the model choice into a spend decision.
After choosing the model id and delivery lane, compare the public pricing dimensions before you put sustained automation traffic behind the key.
Request fieldmodel
Delivery laneDefault, Economy, Coding, Premium, or Spot
Rate sourcePublished model price tables