Closed model API
Onrup vs AWS Bedrock
Should I use Onrup or AWS Bedrock?
Bedrock is the better choice when you need several model vendors behind one AWS contract with governance attached. Onrup is the better choice when you need the customised model to be yours and to run anywhere.
A managed API over several model vendors, with custom-model fine-tuning and provisioned throughput inside AWS.
Where they differ
| Onrup | AWS Bedrock | |
|---|---|---|
| Weight ownership | Download or publish at any time | Custom models stay inside Bedrock |
| Model vendors | Open-weight models only | Several commercial vendors behind one API |
| Serving | Per GPU-second, scale-to-zero available | Provisioned model units per hour |
| Governance | Audit log, scoped keys, spend limits | Native AWS governance and billing |
| Evaluation gate | Blinded, blocking, before deploy | Evaluation features, not blocking |
On price
Custom models are billed per training token plus provisioned model units per hour for serving; no per-GPU-hour rate exists to compare.
Rather than estimate a comparable figure, we leave it blank. Their pricing page, checked 2026-08-06.
The longer answer
Bedrock exists to solve a procurement problem as much as a technical one: several model vendors, one contract, one bill, one set of governance controls. For a large organisation that is a genuinely hard problem and Bedrock solves it well.
The cost is that a custom model made there stays there. Fine-tuning inside Bedrock produces something you can invoke and cannot hold, which is the same trade as any closed platform, wrapped in better governance.
Provisioned throughput is also a different cost shape from GPU-hours. It is predictable and it is committed; scaling to zero between bursts is not part of the model.
If the requirement is vendor choice inside AWS governance, Bedrock. If it is a model you own, us.
Where AWS Bedrock wins
One API across several model vendors, inside an AWS account, with the governance story already written. For an enterprise that needs to switch model providers without switching procurement, that is the whole value proposition.
Choose them if
- You need several model vendors behind one contract
- AWS governance and billing integration is a requirement
- Provisioned throughput suits your traffic better than per-hour GPUs
On ownership
Weights on Onrup are downloadable from every finished run and publishable to a model hub in one call. On AWS Bedrock: Custom models stay inside Bedrock. That is the structural difference between the two products, and it is worth deciding on before you build a dataset rather than after.
Frequently asked questions
Can I export a Bedrock custom model?
No. That is the structural difference, and it is the reason most people evaluating both are evaluating both.
Do you offer provisioned throughput?
The equivalent is an always-warm endpoint with a replica floor — capacity reserved and billed by the hour. There is no separate commitment product.
Can I use both?
Yes. Bedrock for the commercial models, us for the open-weight ones you need to own. Both speak familiar request shapes, so a routing layer is not difficult.
Researching AWS Bedrock alternatives more broadly? →
Last verified 6 August 2026. AWS Bedrock figures come from their own pricing page on the date checked.
Start with the free tier
A magic link creates your account, your tenant and your first API key. No card until you ask for compute.