AvailableService
Fine-tuning Service
Upload examples, get a private adapter.
Managed LoRA, full and preference fine-tuning for the Axiom family, with automatic evaluation against your held-out set before the adapter is promoted to an endpoint.
Specifications
| Methods | LoRA, QLoRA, full SFT, DPO |
|---|---|
| Data | Private, not used for base training |
| Gate | Held-out eval before promotion |
| Output | Downloadable adapter weights |
| Family | AXION Cloud |
| Software stack | FLUX — source-compatible across the whole catalog |
Disclosure
Fine-tuning Service is invented. The numbers above were written to be internally consistent and plausible, not measured. Treat every figure on this site as placeholder content for layout, search and API testing.
Related
Often deployed together
AXION CloudAvailable
Cloud Instances
Accelerators by the second, no minimum term.
- SKUs
- M400, M400X, M200, V90, V40, V20
- Billing
- Per second, 2-minute minimum
ServiceFrom $0.62/hr (V20) to $34.80/hr (8 × M400)
AXION CloudAvailable
Reserved Clusters
Contiguous capacity with a fabric you can benchmark.
- Size
- 32 – 4,096 accelerators
- Terms
- 1 month – 3 years
ServiceFrom $1.94 per accelerator-hour (1 yr)
AXION CloudAvailable
Interruptible Capacity
Cheap accelerators with a 90-second warning.
- Discount
- 62–78% off on-demand
- Warning
- 90 s SIGTERM + hook
ServiceFrom $0.19/hr
AXION CloudAvailable
Serverless Endpoints
Deploy a model, get a URL, pay per token.
- Models
- Catalog + your own weights
- Cold start
- 8.6 s median (warm cache)
ServicePer-token, from $0.04 per M