Prototype operator surface. Benchmark rows come from /v1/benchmark/optimizer (local fixture). Compare routing experiments before changing production weights.
Candidates
0
providers/models in fixture
Workloads
0
prompt-class routes
Avg success
0%
across fixture candidates
Direct API cost
$0
free-tier weighted
Model benchmark matrix
| Model | Provider | Cost /1M | p50 | p90 | Success | Quality | Cooldown | Status | Score |
|---|
Recommended route by workload
Experiment ledger
Fixture entries plus browser-local snapshots for before/after comparisons.
| When | Label | Change | Success | Latency | Cost /1k | Fallback |
|---|