Function-Calling Benchmark Tutorials#
These tutorials follow complete BFCL journeys. Start with the bundled tiny pack to learn the artifact flow, run the credential-free assisted-authoring demonstration to observe every review and publication gate, then use the manual lifecycle when you are ready to publish and evaluate a pack of your own.
Build Your First Benchmark
Run the bundled tiny_oracle_pack, inspect its validation report, benchmark rows, and
publication manifest, without a model endpoint.
Run Assisted Authoring End to End
Watch source intake, certification, drafting, human-review boundaries, Gold validation, publication, and evaluation in one credential-free demonstration.
Publish and Evaluate a Manual Pack
Take a reviewed Oracle Pack through validation, publication, independent-candidate preflight, live evaluation, and artifact inspection.