Specifications
| Routing | Learned difficulty classifier |
|---|---|
| Failover | Cross-provider, per route |
| Savings | 31–58% observed in pilots |
| Shadow mode | Compare before cutover |
| Family | Inference & Serving |
| Software stack | FLUX — source-compatible across the whole catalog |
Disclosure
Broker Model Router 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
Inference & ServingAvailable
Kiln Compiler
Take a trained model, get a fast engine.
- Inputs
- PyTorch, ONNX, JAX export, GGUF
- Quantization
- FP8, FP4, INT8, INT4 + accuracy report
SDKFree
Inference & ServingAvailable
Conductor Inference Server
One server, every framework, real batching.
- Backends
- Kiln, PyTorch, ONNX, vLLM, XGBoost
- Batching
- Continuous + prefill/decode split
PlatformFree (open source)
Inference & ServingAvailable
AXION Cells
Pre-built model containers with one API each.
- Catalog
- 310 model containers
- Deploy targets
- K8s, bare metal, edge, cloud
Platform$4,500 per accelerator per year
Inference & ServingAvailable
Slipstream Distributed Serving
Disaggregated inference across a whole rack.
- Architecture
- Disaggregated prefill/decode
- KV transfer
- HyperLink and RDMA
PlatformIncluded with Cells