Developer PreviewPublic result artifact

Leaderboard

A future ranking surface for versioned Tutor capability results. Until the corpus is reproducible and calibrated, this page stays explicit about what is not known.

No calibrated public model runs yet. Public model results are unavailable. Human calibration (P5) has not started, and Judge-vs-human and statistical validation are not completed.

Leaderboard coming soon

No real model results are checked into the public artifact. Synthetic demonstrations are not presented as rankings.

Future public model leaderboard
RankModelOverallCorrectnessDiagnosisGuidanceAdaptationActionabilityStatus
No public model rows are available yet.
No public model rows are available yet.

Tutor capability score

CorrectnessDiagnosisGuidanceAdaptationActionability

Operational signals

CriticalFailureRateAnswerLeakageRateLatencyMsTokensCost

Traceability fields

DatasetVersionGenerationSpecIdGenerationSpecVersionPromptVersionPromptSha256Runs

The future table will show: Correctness, Diagnosis, Guidance, Adaptation, Actionability. Operational fields include CriticalFailureRate, AnswerLeakageRate, LatencyMs, Tokens, Cost. Results from different generation profiles are separate cohorts and are not silently mixed.

Future filters

Designed for comparison without hiding context

The v0.1 shell leaves room for benchmark, prompt, subject, provider, model, and metric sorting without implementing a complex filter engine yet.

Version
Benchmark version · prompt version · dataset version
Scope
Subject · model provider · model
Sort
Overall · correctness · diagnosis · guidance · adaptation · actionability · cost · latency