Before launch
What still has to happen
- Freeze prompt construction and adjudication rules.
- Secure production-equivalent access for the first models.
- Run replication checks and document access sources.
- Publish limitations with the first table.
VetoBench / Leaderboard
No model has been ranked yet.
The leaderboard will open after the Pilot 0.1 protocol is frozen, evaluation access is recorded and the first runs pass adjudication.
ProtocolDraft
Evaluated models0
Published runs0
Next milestonePilot freeze
ModelBenign completionVeto rateStabilityExtensionEvaluated
Awaiting valid pilot results
Rows will not be populated with estimates, vendor claims or hand-entered marketing scores.