ModelSkyline day-one provisional frontier
This is a separate launch-day view. It is not a mature Skyline quality ranking and is not an input to model selection.
- Catalog evidence: model identity, availability, pricing, context and published capabilities.
- Operational evidence: aggregator telemetry, where available, labelled separately.
- Published evaluations: benchmark values are included only with an explicit source, benchmark name and methodology, and remain provenance-labelled.
- Eligibility: every record is marked
mature_evaluation_eligible: false and selection_eligible: false.
Download the machine-readable provisional-frontier.json artifact.
The mature snapshot remains at Frontier snapshots and summary-v2.json.