Sources
0
Comparative Workbench
BitterBench is becoming a real comparative inference surface.
This workbench now supports durable comparative ingest, multi-plane source summaries, filterable benchmark packets, and direct run inspection for the selected comparison set.
Current State
Comparison drilldown live
BitterMill exports and external-provider packets can now land in one contract and be inspected as comparison sets rather than flat logs.
Runs
0
Completed
0
Latest
n/a
Filters
Narrow the benchmark corpus by plane, provider, source, workload, or model.
0 active
Comparison Sets
Select a benchmark packet to inspect its sources and underlying runs.
No comparison sets yet. Ingest BitterMill exports or demo packets to start building a comparative corpus.
Select a comparison set to inspect it.
Runs In Selection
Inspect the underlying benchmark records for the active comparison.
0 of 0 loaded
No runs available for the current selection.
Select a run to inspect its timings and metadata.
Sources
BitterMill cells, routers, and cloud providers should all land in the same source model.
No sources ingested yet. POST a comparative packet or load the demo samples to populate the workbench.
Latest Runs
Recent benchmark records across the currently filtered source set.
No run records ingested yet. Start with `api/bin/load_demo_benchmark_data` or post a comparative payload into the ingest endpoint.