Measurements
Benchmarks
Charts on this page come only from benchmark-v1 reports committed in source repositories. A generator writes each report, and its raw data is checksummed. The build re-checks the checksum, so a hand-edited number fails CI.
Nothing published yet
No benchmark reports are published yet
None of the pinned source releases contains a benchmark-v1 report yet. The first planned one is a k6 load test of distributed, token-aware rate limiting. It will appear here, with its raw data and environment, once it exists.
BoundedCode publishes task-outcome validation reports. Those are evaluations, not performance benchmarks, so they are linked as reports from the case study instead. The report format is documented in the benchmark-v1 schema.