Pricing built around your workload.
Pricing depends on the models you run, expected volume, throughput, and where the workload needs to run. We quote usage, dedicated capacity, reserved throughput, and custom deployments against your requirements.
What shapes your price
Confirmed during onboarding, before production
Model APIs

Ready-to-use models, charged on usage. Run in a shared pool or as a dedicated deployment.

What you pay for
Basis
Availability
Usage
Per million tokens
On request
Reserved throughput
Agreed rate
On request
Custom model inference

Your weights or container, optimised and scaled by us. Dedicated deployment on our GPUs or in your environment.

What you pay for
Basis
Availability
Reserved GPUs
Hourly, metered per minute
On request
Platform services
Added to GPU cost
On request
What changes the price

The same workload can price differently depending on commitment, volume, and where it has to run.

Factor
Effect
Discussed at
Contract length and volume
Discounts available
Onboarding
Processing location requirements
Increases price
Onboarding
Batch and sandboxes

Large volumes with an agreed completion deadline, and isolated environments for agents, are priced separately.

Service
Basis
Availability
Batch processing
Separate rate
On request
Sandboxes
CPU/RAM time or per session
On request
Tell us what you're running.
Share the models, expected volume, latency targets, and any placement requirements. We will come back with pricing for the right serving profile.
Pages
HomeModelsPricingBlogBook a demo
Platform
EnterpriseSecurity
Legals
Privacy PolicyTerms of Service
Contact
hello@aleroy.aiTalk to sales
@ 2026 Aleroy. Not your typical inference network.