{ "schemaVersion": 1, "snapshotDate": "2026-09-30", "collectedAtUtc": "2026-09-30T21:52:50Z", "collectionStartedAtUtc": "2026-09-30T21:47:59Z", "observationWindow": { "startInclusive": "2026-09-24T00:00:00Z", "endExclusive": "2026-09-30T00:00:00Z", "timeZone": "UTC", "dateField": "Job-record creation timestamp" }, "productionPerformanceGroup": { "minimumVersion": "2026-09-23", "observedVersions": [ "2026-09-23" ], "versionFilter": "Whole YYYY-MM-DD release tags on the production inference endpoint" }, "summary": { "storedJobs": { "total": 5490, "outcomes": { "COMPLETED": 5452, "FAILED": 38, "CANCELED": 0, "PENDING": 0, "UNKNOWN": 0 }, "analyticsMissing": 0, "statusMismatches": 0, "scope": "Stored jobs created during the fixed UTC window, regardless of available version metadata. Requests rejected before job-record creation are absent." }, "latestPerformanceGroup": { "total": 5487, "outcomes": { "COMPLETED": 5450, "FAILED": 37, "CANCELED": 0, "PENDING": 0, "UNKNOWN": 0 }, "minimumObservedVersion": "2026-09-23", "maximumObservedVersion": "2026-09-23", "completedAudioDurationMissing": 0, "failedAudioDurationMissing": 37, "englishRequestedCompleted": 1, "verifiedEnglishAudio": null, "verifiedEnglishAudioNote": "Detected corpus language is not stored; a requested language hint does not verify the audio language.", "firstWorkerRunCompleted": 838, "laterWorkerRunCompleted": 4612, "unknownRunCountCompleted": 0 }, "excludedFromLatestPerformanceGroup": [ { "status": "COMPLETED", "reason": "older_date_tag", "count": 2 }, { "status": "FAILED", "reason": "version_unavailable", "count": 1 } ], "reconciliation": { "storedJobs": 5490, "analyticsMissing": 0, "statusMismatches": 0, "startTimeMismatches": 0, "analyticsWithoutStoredJob": 0, "latestCompletedFromJobs": 5450, "latestFailedFromJobs": 37 } }, "metricDefinitions": { "job_elapsed_seconds": { "label": "Server-side job elapsed time", "unit": "seconds", "start": "Job-record creation after URL prefetch and inference dispatch.", "end": "Timestamp captured in the completion handler after updating job status and before persisting analytics, storing the transcript, formatting delivery or sending the customer webhook.", "excludes": [ "Initial request handling and URL prefetch before the job record exists", "Dispatch work before job-record creation", "Transcript storage and customer delivery after the measured timestamp", "Client polling, network receipt and result download" ], "note": "Callback retries can overwrite analytics with a later elapsed measurement. This is not a client submission-to-result timer." }, "provider_delay_seconds": { "label": "Inference delay", "unit": "seconds", "definition": "Delay reported by the inference platform, converted from milliseconds. Includes platform waiting and startup effects. It is not a separate client timer." }, "provider_execution_seconds": { "label": "Inference execution", "unit": "seconds", "definition": "Execution reported by the inference platform, converted from milliseconds. This includes work performed inside the inference job rather than only speech recognition." }, "audio_seconds": { "label": "Recorded audio duration", "unit": "seconds", "definition": "Duration returned by the completed job. Primary presentation should convert seconds to minutes." }, "effectiveSpeed": { "label": "Effective processing speed", "unit": "times audio duration", "definition": "Sum of audio seconds divided by sum of server-side job elapsed seconds for the same completed observations with both values positive. This is a ratio of sums and does not measure concurrent endpoint throughput." } }, "cohorts": [ { "cohort": "about_5_minutes__all_settings", "languageScope": "all_languages", "outcomes": { "analytics_rows": 145, "completed": 145, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 145, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 145, "samples": 145, "missing": 0, "invalid": 0, "minimum": 270.108, "mean": 299.38411724137933, "p50": 300.42, "p90": 323.82, "p95": 325.908 }, "job_elapsed_seconds": { "completed": 145, "samples": 145, "missing": 0, "invalid": 0, "minimum": 3.253, "mean": 19.130413793103447, "p50": 9.739, "p90": 26.931, "p95": 41.81 }, "provider_delay_seconds": { "completed": 145, "samples": 145, "missing": 0, "invalid": 0, "minimum": 0.461, "mean": 11.309813793103448, "p50": 1.217, "p90": 14.005, "p95": 30.735 }, "provider_execution_seconds": { "completed": 145, "samples": 145, "missing": 0, "invalid": 0, "minimum": 1.415, "mean": 7.240468965517241, "p50": 6.547, "p90": 11.65, "p95": 13.235 } }, "effectiveSpeed": { "samples": 145, "pairedAudioSeconds": 43410.697, "pairedJobElapsedSeconds": 2773.91, "ratioOfSums": 15.649641480797865 }, "tailInterpretation": "Observed nearest-rank sample percentiles; no population guarantee." }, { "cohort": "about_5_minutes__all_settings", "languageScope": "english_requested", "outcomes": { "analytics_rows": 0, "completed": 0, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 0, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "job_elapsed_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_delay_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_execution_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null } }, "effectiveSpeed": { "samples": 0, "pairedAudioSeconds": 0, "pairedJobElapsedSeconds": 0, "ratioOfSums": null }, "tailInterpretation": "No observations." }, { "cohort": "about_5_minutes__base", "languageScope": "all_languages", "outcomes": { "analytics_rows": 133, "completed": 133, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 133, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 133, "samples": 133, "missing": 0, "invalid": 0, "minimum": 270.108, "mean": 298.23347368421054, "p50": 299.376, "p90": 320.112, "p95": 325.584 }, "job_elapsed_seconds": { "completed": 133, "samples": 133, "missing": 0, "invalid": 0, "minimum": 3.464, "mean": 18.306323308270677, "p50": 9.739, "p90": 26.931, "p95": 41.81 }, "provider_delay_seconds": { "completed": 133, "samples": 133, "missing": 0, "invalid": 0, "minimum": 0.461, "mean": 10.524706766917294, "p50": 1.561, "p90": 14.005, "p95": 30.735 }, "provider_execution_seconds": { "completed": 133, "samples": 133, "missing": 0, "invalid": 0, "minimum": 1.415, "mean": 7.178872180451128, "p50": 6.451, "p90": 11.561, "p95": 13.006 } }, "effectiveSpeed": { "samples": 133, "pairedAudioSeconds": 39665.052, "pairedJobElapsedSeconds": 2434.741, "ratioOfSums": 16.29128190637115 }, "tailInterpretation": "Observed nearest-rank sample percentiles; no population guarantee." }, { "cohort": "about_5_minutes__base", "languageScope": "english_requested", "outcomes": { "analytics_rows": 0, "completed": 0, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 0, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "job_elapsed_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_delay_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_execution_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null } }, "effectiveSpeed": { "samples": 0, "pairedAudioSeconds": 0, "pairedJobElapsedSeconds": 0, "ratioOfSums": null }, "tailInterpretation": "No observations." }, { "cohort": "about_5_minutes__diarization", "languageScope": "all_languages", "outcomes": { "analytics_rows": 0, "completed": 0, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 0, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "job_elapsed_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_delay_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_execution_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null } }, "effectiveSpeed": { "samples": 0, "pairedAudioSeconds": 0, "pairedJobElapsedSeconds": 0, "ratioOfSums": null }, "tailInterpretation": "No observations." }, { "cohort": "about_5_minutes__diarization", "languageScope": "english_requested", "outcomes": { "analytics_rows": 0, "completed": 0, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 0, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "job_elapsed_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_delay_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_execution_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null } }, "effectiveSpeed": { "samples": 0, "pairedAudioSeconds": 0, "pairedJobElapsedSeconds": 0, "ratioOfSums": null }, "tailInterpretation": "No observations." }, { "cohort": "about_60_minutes__all_settings", "languageScope": "all_languages", "outcomes": { "analytics_rows": 62, "completed": 62, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 1, "automatic_or_unspecified": 61, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 62, "samples": 62, "missing": 0, "invalid": 0, "minimum": 3336.96, "mean": 3570.756725806452, "p50": 3525.077, "p90": 3853.505, "p95": 3886.289 }, "job_elapsed_seconds": { "completed": 62, "samples": 62, "missing": 0, "invalid": 0, "minimum": 19.809, "mean": 56.09933870967742, "p50": 48.182, "p90": 84.64, "p95": 104.939 }, "provider_delay_seconds": { "completed": 62, "samples": 62, "missing": 0, "invalid": 0, "minimum": 0.016, "mean": 3.7060645161290324, "p50": 0.855, "p90": 10.586, "p95": 12.333 }, "provider_execution_seconds": { "completed": 62, "samples": 62, "missing": 0, "invalid": 0, "minimum": 15.794, "mean": 51.945016129032254, "p50": 44.834, "p90": 83.756, "p95": 104.163 } }, "effectiveSpeed": { "samples": 62, "pairedAudioSeconds": 221386.91700000002, "pairedJobElapsedSeconds": 3478.159, "ratioOfSums": 63.65060280452964 }, "tailInterpretation": "Observed nearest-rank sample percentiles; no population guarantee." }, { "cohort": "about_60_minutes__all_settings", "languageScope": "english_requested", "outcomes": { "analytics_rows": 1, "completed": 1, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 1, "automatic_or_unspecified": 0, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 1, "samples": 1, "missing": 0, "invalid": 0, "minimum": 3494.507, "mean": 3494.507, "p50": 3494.507, "p90": 3494.507, "p95": 3494.507 }, "job_elapsed_seconds": { "completed": 1, "samples": 1, "missing": 0, "invalid": 0, "minimum": 30.625, "mean": 30.625, "p50": 30.625, "p90": 30.625, "p95": 30.625 }, "provider_delay_seconds": { "completed": 1, "samples": 1, "missing": 0, "invalid": 0, "minimum": 14.612, "mean": 14.612, "p50": 14.612, "p90": 14.612, "p95": 14.612 }, "provider_execution_seconds": { "completed": 1, "samples": 1, "missing": 0, "invalid": 0, "minimum": 15.794, "mean": 15.794, "p50": 15.794, "p90": 15.794, "p95": 15.794 } }, "effectiveSpeed": { "samples": 1, "pairedAudioSeconds": 3494.507, "pairedJobElapsedSeconds": 30.625, "ratioOfSums": 114.10635102040817 }, "tailInterpretation": "Small sample counts do not establish a stable population tail." }, { "cohort": "about_60_minutes__base", "languageScope": "all_languages", "outcomes": { "analytics_rows": 8, "completed": 8, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 8, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 8, "samples": 8, "missing": 0, "invalid": 0, "minimum": 3358.62, "mean": 3594.6405, "p50": 3426.84, "p90": 3893.76, "p95": 3893.76 }, "job_elapsed_seconds": { "completed": 8, "samples": 8, "missing": 0, "invalid": 0, "minimum": 27.497, "mean": 43.51875, "p50": 32.485, "p90": 70.078, "p95": 70.078 }, "provider_delay_seconds": { "completed": 8, "samples": 8, "missing": 0, "invalid": 0, "minimum": 0.52, "mean": 6.184, "p50": 1.078, "p90": 24.267, "p95": 24.267 }, "provider_execution_seconds": { "completed": 8, "samples": 8, "missing": 0, "invalid": 0, "minimum": 20.552, "mean": 36.9695, "p50": 27.073, "p90": 62.323, "p95": 62.323 } }, "effectiveSpeed": { "samples": 8, "pairedAudioSeconds": 28757.124, "pairedJobElapsedSeconds": 348.15, "ratioOfSums": 82.59981042654029 }, "tailInterpretation": "Small sample counts do not establish a stable population tail." }, { "cohort": "about_60_minutes__base", "languageScope": "english_requested", "outcomes": { "analytics_rows": 0, "completed": 0, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 0, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "job_elapsed_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_delay_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_execution_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null } }, "effectiveSpeed": { "samples": 0, "pairedAudioSeconds": 0, "pairedJobElapsedSeconds": 0, "ratioOfSums": null }, "tailInterpretation": "No observations." }, { "cohort": "about_60_minutes__diarization", "languageScope": "all_languages", "outcomes": { "analytics_rows": 0, "completed": 0, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 0, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "job_elapsed_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_delay_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_execution_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null } }, "effectiveSpeed": { "samples": 0, "pairedAudioSeconds": 0, "pairedJobElapsedSeconds": 0, "ratioOfSums": null }, "tailInterpretation": "No observations." }, { "cohort": "about_60_minutes__diarization", "languageScope": "english_requested", "outcomes": { "analytics_rows": 0, "completed": 0, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 0, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "job_elapsed_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_delay_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_execution_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null } }, "effectiveSpeed": { "samples": 0, "pairedAudioSeconds": 0, "pairedJobElapsedSeconds": 0, "ratioOfSums": null }, "tailInterpretation": "No observations." }, { "cohort": "all_durations__all_settings", "languageScope": "all_languages", "outcomes": { "analytics_rows": 5487, "completed": 5450, "failed": 37, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 1, "automatic_or_unspecified": 5486, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 5450, "samples": 5450, "missing": 0, "invalid": 0, "minimum": 6.074, "mean": 508.84691064220186, "p50": 70.56, "p90": 1606.925, "p95": 2704.032 }, "job_elapsed_seconds": { "completed": 5450, "samples": 5450, "missing": 0, "invalid": 0, "minimum": 1.197, "mean": 19.364855412844037, "p50": 9.292, "p90": 36.811, "p95": 66.552 }, "provider_delay_seconds": { "completed": 5450, "samples": 5450, "missing": 0, "invalid": 0, "minimum": 0.016, "mean": 9.383445504587156, "p50": 2.023, "p90": 13.925, "p95": 25.368 }, "provider_execution_seconds": { "completed": 5450, "samples": 5450, "missing": 0, "invalid": 0, "minimum": 0.523, "mean": 9.632370642201835, "p50": 4.186, "p90": 20.038, "p95": 34.852 } }, "effectiveSpeed": { "samples": 5450, "pairedAudioSeconds": 2773215.663, "pairedJobElapsedSeconds": 105538.462, "ratioOfSums": 26.276824680276277 }, "tailInterpretation": "Observed nearest-rank sample percentiles; no population guarantee." }, { "cohort": "all_durations__all_settings", "languageScope": "english_requested", "outcomes": { "analytics_rows": 1, "completed": 1, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 1, "automatic_or_unspecified": 0, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 1, "samples": 1, "missing": 0, "invalid": 0, "minimum": 3494.507, "mean": 3494.507, "p50": 3494.507, "p90": 3494.507, "p95": 3494.507 }, "job_elapsed_seconds": { "completed": 1, "samples": 1, "missing": 0, "invalid": 0, "minimum": 30.625, "mean": 30.625, "p50": 30.625, "p90": 30.625, "p95": 30.625 }, "provider_delay_seconds": { "completed": 1, "samples": 1, "missing": 0, "invalid": 0, "minimum": 14.612, "mean": 14.612, "p50": 14.612, "p90": 14.612, "p95": 14.612 }, "provider_execution_seconds": { "completed": 1, "samples": 1, "missing": 0, "invalid": 0, "minimum": 15.794, "mean": 15.794, "p50": 15.794, "p90": 15.794, "p95": 15.794 } }, "effectiveSpeed": { "samples": 1, "pairedAudioSeconds": 3494.507, "pairedJobElapsedSeconds": 30.625, "ratioOfSums": 114.10635102040817 }, "tailInterpretation": "Small sample counts do not establish a stable population tail." }, { "cohort": "all_durations__base", "languageScope": "all_languages", "outcomes": { "analytics_rows": 4922, "completed": 4885, "failed": 37, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 4922, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 4885, "samples": 4885, "missing": 0, "invalid": 0, "minimum": 6.074, "mean": 258.6603604912999, "p50": 54.612, "p90": 804.564, "p95": 1351.116 }, "job_elapsed_seconds": { "completed": 4885, "samples": 4885, "missing": 0, "invalid": 0, "minimum": 1.197, "mean": 16.300172364380757, "p50": 8.297, "p90": 26.547, "p95": 41.317 }, "provider_delay_seconds": { "completed": 4885, "samples": 4885, "missing": 0, "invalid": 0, "minimum": 0.018, "mean": 9.98176724667349, "p50": 2.687, "p90": 14.825, "p95": 28.091 }, "provider_execution_seconds": { "completed": 4885, "samples": 4885, "missing": 0, "invalid": 0, "minimum": 0.523, "mean": 5.987442169907881, "p50": 3.72, "p90": 13.039, "p95": 18.443 } }, "effectiveSpeed": { "samples": 4885, "pairedAudioSeconds": 1263555.861, "pairedJobElapsedSeconds": 79626.342, "ratioOfSums": 15.86856597029159 }, "tailInterpretation": "Observed nearest-rank sample percentiles; no population guarantee." }, { "cohort": "all_durations__base", "languageScope": "english_requested", "outcomes": { "analytics_rows": 0, "completed": 0, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 0, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "job_elapsed_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_delay_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_execution_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null } }, "effectiveSpeed": { "samples": 0, "pairedAudioSeconds": 0, "pairedJobElapsedSeconds": 0, "ratioOfSums": null }, "tailInterpretation": "No observations." }, { "cohort": "all_durations__diarization", "languageScope": "all_languages", "outcomes": { "analytics_rows": 0, "completed": 0, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 0, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "job_elapsed_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_delay_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_execution_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null } }, "effectiveSpeed": { "samples": 0, "pairedAudioSeconds": 0, "pairedJobElapsedSeconds": 0, "ratioOfSums": null }, "tailInterpretation": "No observations." }, { "cohort": "all_durations__diarization", "languageScope": "english_requested", "outcomes": { "analytics_rows": 0, "completed": 0, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 0, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "job_elapsed_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_delay_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_execution_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null } }, "effectiveSpeed": { "samples": 0, "pairedAudioSeconds": 0, "pairedJobElapsedSeconds": 0, "ratioOfSums": null }, "tailInterpretation": "No observations." } ], "durationBuckets": [ { "cohort": "0_to_1_minutes", "languageScope": "all_languages", "outcomes": { "analytics_rows": 2560, "completed": 2560, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 2560, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 2560, "samples": 2560, "missing": 0, "invalid": 0, "minimum": 6.074, "mean": 29.032425390625, "p50": 28.224, "p90": 48.672, "p95": 53.856 }, "job_elapsed_seconds": { "completed": 2560, "samples": 2560, "missing": 0, "invalid": 0, "minimum": 1.197, "mean": 13.162730859375, "p50": 5.64, "p90": 18.514, "p95": 29.87 }, "provider_delay_seconds": { "completed": 2560, "samples": 2560, "missing": 0, "invalid": 0, "minimum": 0.122, "mean": 9.8890609375, "p50": 2.366, "p90": 13.968, "p95": 25.879 }, "provider_execution_seconds": { "completed": 2560, "samples": 2560, "missing": 0, "invalid": 0, "minimum": 0.523, "mean": 2.974229296875, "p50": 2.39, "p90": 5.727, "p95": 6.876 } }, "effectiveSpeed": { "samples": 2560, "pairedAudioSeconds": 74323.009, "pairedJobElapsedSeconds": 33696.591, "ratioOfSums": 2.205653652026699 }, "tailInterpretation": "Observed nearest-rank sample percentiles; no population guarantee." }, { "cohort": "12_to_24_hours", "languageScope": "all_languages", "outcomes": { "analytics_rows": 0, "completed": 0, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 0, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "job_elapsed_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_delay_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_execution_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null } }, "effectiveSpeed": { "samples": 0, "pairedAudioSeconds": 0, "pairedJobElapsedSeconds": 0, "ratioOfSums": null }, "tailInterpretation": "No observations." }, { "cohort": "15_to_30_minutes", "languageScope": "all_languages", "outcomes": { "analytics_rows": 402, "completed": 402, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 402, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 402, "samples": 402, "missing": 0, "invalid": 0, "minimum": 901.512, "mean": 1292.8493606965176, "p50": 1252.656, "p90": 1682.604, "p95": 1732.464 }, "job_elapsed_seconds": { "completed": 402, "samples": 402, "missing": 0, "invalid": 0, "minimum": 3.187, "mean": 26.96741791044776, "p50": 20.901, "p90": 40.621, "p95": 62.957 }, "provider_delay_seconds": { "completed": 402, "samples": 402, "missing": 0, "invalid": 0, "minimum": 0.019, "mean": 7.534271144278607, "p50": 2.827, "p90": 12.509, "p95": 22.417 }, "provider_execution_seconds": { "completed": 402, "samples": 402, "missing": 0, "invalid": 0, "minimum": 1.473, "mean": 19.037320895522388, "p50": 15.998, "p90": 26.554, "p95": 36.362 } }, "effectiveSpeed": { "samples": 402, "pairedAudioSeconds": 519725.443, "pairedJobElapsedSeconds": 10840.902, "ratioOfSums": 47.94116236822361 }, "tailInterpretation": "Observed nearest-rank sample percentiles; no population guarantee." }, { "cohort": "1_to_2_hours", "languageScope": "all_languages", "outcomes": { "analytics_rows": 143, "completed": 143, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 143, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 143, "samples": 143, "missing": 0, "invalid": 0, "minimum": 3609.495, "mean": 4788.880755244755, "p50": 4607.425, "p90": 6094.179, "p95": 6540.278 }, "job_elapsed_seconds": { "completed": 143, "samples": 143, "missing": 0, "invalid": 0, "minimum": 18.63, "mean": 73.7018951048951, "p50": 62.003, "p90": 122.719, "p95": 176.694 }, "provider_delay_seconds": { "completed": 143, "samples": 143, "missing": 0, "invalid": 0, "minimum": 0.016, "mean": 4.469832167832168, "p50": 0.865, "p90": 8.623, "p95": 10.698 }, "provider_execution_seconds": { "completed": 143, "samples": 143, "missing": 0, "invalid": 0, "minimum": 17.891, "mean": 68.65253846153846, "p50": 57.884, "p90": 117.766, "p95": 158.7 } }, "effectiveSpeed": { "samples": 143, "pairedAudioSeconds": 684809.948, "pairedJobElapsedSeconds": 10539.371, "ratioOfSums": 64.97635845630636 }, "tailInterpretation": "Observed nearest-rank sample percentiles; no population guarantee." }, { "cohort": "1_to_5_minutes", "languageScope": "all_languages", "outcomes": { "analytics_rows": 1346, "completed": 1346, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 1346, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 1346, "samples": 1346, "missing": 0, "invalid": 0, "minimum": 60.048, "mean": 139.39955497771174, "p50": 122.22, "p90": 246.204, "p95": 271.368 }, "job_elapsed_seconds": { "completed": 1346, "samples": 1346, "missing": 0, "invalid": 0, "minimum": 1.772, "mean": 16.063885586924222, "p50": 8.118, "p90": 23.544, "p95": 39.54 }, "provider_delay_seconds": { "completed": 1346, "samples": 1346, "missing": 0, "invalid": 0, "minimum": 0.018, "mean": 10.392607726597326, "p50": 2.506, "p90": 15.324, "p95": 33.113 }, "provider_execution_seconds": { "completed": 1346, "samples": 1346, "missing": 0, "invalid": 0, "minimum": 0.698, "mean": 5.305604754829123, "p50": 4.673, "p90": 8.409, "p95": 10.407 } }, "effectiveSpeed": { "samples": 1346, "pairedAudioSeconds": 187631.801, "pairedJobElapsedSeconds": 21621.99, "ratioOfSums": 8.677822947841525 }, "tailInterpretation": "Observed nearest-rank sample percentiles; no population guarantee." }, { "cohort": "24_plus_hours", "languageScope": "all_languages", "outcomes": { "analytics_rows": 0, "completed": 0, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 0, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "job_elapsed_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_delay_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_execution_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null } }, "effectiveSpeed": { "samples": 0, "pairedAudioSeconds": 0, "pairedJobElapsedSeconds": 0, "ratioOfSums": null }, "tailInterpretation": "No observations." }, { "cohort": "2_to_4_hours", "languageScope": "all_languages", "outcomes": { "analytics_rows": 13, "completed": 13, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 13, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 13, "samples": 13, "missing": 0, "invalid": 0, "minimum": 7305.971, "mean": 8742.843692307692, "p50": 8785.227, "p90": 9813.864, "p95": 10572.437 }, "job_elapsed_seconds": { "completed": 13, "samples": 13, "missing": 0, "invalid": 0, "minimum": 85.324, "mean": 178.5086923076923, "p50": 111.183, "p90": 289.546, "p95": 591.701 }, "provider_delay_seconds": { "completed": 13, "samples": 13, "missing": 0, "invalid": 0, "minimum": 0.477, "mean": 3.5502307692307693, "p50": 0.856, "p90": 9.753, "p95": 11.012 }, "provider_execution_seconds": { "completed": 13, "samples": 13, "missing": 0, "invalid": 0, "minimum": 84.472, "mean": 174.20046153846155, "p50": 106.265, "p90": 288.785, "p95": 589.62 } }, "effectiveSpeed": { "samples": 13, "pairedAudioSeconds": 113656.968, "pairedJobElapsedSeconds": 2320.613, "ratioOfSums": 48.977131473451195 }, "tailInterpretation": "Small sample counts do not establish a stable population tail." }, { "cohort": "30_to_60_minutes", "languageScope": "all_languages", "outcomes": { "analytics_rows": 326, "completed": 326, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 1, "automatic_or_unspecified": 325, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 326, "samples": 326, "missing": 0, "invalid": 0, "minimum": 1803.492, "mean": 2524.9517239263805, "p50": 2427.084, "p90": 3368.359, "p95": 3424.8 }, "job_elapsed_seconds": { "completed": 326, "samples": 326, "missing": 0, "invalid": 0, "minimum": 10.521, "mean": 39.49884355828221, "p50": 31.785, "p90": 63.25, "p95": 87.602 }, "provider_delay_seconds": { "completed": 326, "samples": 326, "missing": 0, "invalid": 0, "minimum": 0.017, "mean": 5.595070552147239, "p50": 1.012, "p90": 10.741, "p95": 16.402 }, "provider_execution_seconds": { "completed": 326, "samples": 326, "missing": 0, "invalid": 0, "minimum": 4.318, "mean": 33.417444785276075, "p50": 26.879, "p90": 55.601, "p95": 66.905 } }, "effectiveSpeed": { "samples": 326, "pairedAudioSeconds": 823134.262, "pairedJobElapsedSeconds": 12876.623, "ratioOfSums": 63.92469997762612 }, "tailInterpretation": "Observed nearest-rank sample percentiles; no population guarantee." }, { "cohort": "4_to_8_hours", "languageScope": "all_languages", "outcomes": { "analytics_rows": 1, "completed": 1, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 1, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 1, "samples": 1, "missing": 0, "invalid": 0, "minimum": 19531.691, "mean": 19531.691, "p50": 19531.691, "p90": 19531.691, "p95": 19531.691 }, "job_elapsed_seconds": { "completed": 1, "samples": 1, "missing": 0, "invalid": 0, "minimum": 153.84, "mean": 153.84, "p50": 153.84, "p90": 153.84, "p95": 153.84 }, "provider_delay_seconds": { "completed": 1, "samples": 1, "missing": 0, "invalid": 0, "minimum": 0.967, "mean": 0.967, "p50": 0.967, "p90": 0.967, "p95": 0.967 }, "provider_execution_seconds": { "completed": 1, "samples": 1, "missing": 0, "invalid": 0, "minimum": 151.318, "mean": 151.318, "p50": 151.318, "p90": 151.318, "p95": 151.318 } }, "effectiveSpeed": { "samples": 1, "pairedAudioSeconds": 19531.691, "pairedJobElapsedSeconds": 153.84, "ratioOfSums": 126.9610699427977 }, "tailInterpretation": "Small sample counts do not establish a stable population tail." }, { "cohort": "5_to_15_minutes", "languageScope": "all_languages", "outcomes": { "analytics_rows": 659, "completed": 659, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 659, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 659, "samples": 659, "missing": 0, "invalid": 0, "minimum": 300.42, "mean": 531.7185751138088, "p50": 483.012, "p90": 811.62, "p95": 860.076 }, "job_elapsed_seconds": { "completed": 659, "samples": 659, "missing": 0, "invalid": 0, "minimum": 2.411, "mean": 20.468182094081943, "p50": 13.496, "p90": 30.836, "p95": 45.7 }, "provider_delay_seconds": { "completed": 659, "samples": 659, "missing": 0, "invalid": 0, "minimum": 0.016, "mean": 9.554257966616085, "p50": 2.593, "p90": 15.021, "p95": 26.664 }, "provider_execution_seconds": { "completed": 659, "samples": 659, "missing": 0, "invalid": 0, "minimum": 1.518, "mean": 10.562552352048558, "p50": 9.06, "p90": 17.4, "p95": 21.46 } }, "effectiveSpeed": { "samples": 659, "pairedAudioSeconds": 350402.541, "pairedJobElapsedSeconds": 13488.532, "ratioOfSums": 25.977811447531877 }, "tailInterpretation": "Observed nearest-rank sample percentiles; no population guarantee." }, { "cohort": "8_to_12_hours", "languageScope": "all_languages", "outcomes": { "analytics_rows": 0, "completed": 0, "failed": 0, "canceled": 0, "pending": 0, "unknown": 0 }, "languageCoverage": { "english_requested": 0, "automatic_or_unspecified": 0, "other_language_requested": 0 }, "metrics": { "audio_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "job_elapsed_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_delay_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null }, "provider_execution_seconds": { "completed": 0, "samples": 0, "missing": 0, "invalid": 0, "minimum": null, "mean": null, "p50": null, "p90": null, "p95": null } }, "effectiveSpeed": { "samples": 0, "pairedAudioSeconds": 0, "pairedJobElapsedSeconds": 0, "ratioOfSums": null }, "tailInterpretation": "No observations." } ], "recordedComputeCharges": { "label": "Recorded compute charges", "amountUsd": 17.473549147056474, "currency": "USD", "source": "Stored hourly billing records copied from the billing API without a modeled rate.", "window": "2026-09-24 through 2026-09-29 UTC", "storedHourlyRecords": 142, "configuredEndpoints": 2, "calendarDaysWithRecords": 6, "hoursWithStoredRecords": 134, "hoursInWindow": 144, "hoursWithoutStoredRecords": 10, "recordedBilledSeconds": 91015.981, "dailyRecords": [ { "day_utc": "2026-09-24", "stored_bills": 25, "hours_with_stored_bills": 24, "recorded_amount_usd": 5.0400230169907445, "unavailable_amounts": 0, "negative_amounts": 0 }, { "day_utc": "2026-09-25", "stored_bills": 26, "hours_with_stored_bills": 24, "recorded_amount_usd": 2.8396947298278974, "unavailable_amounts": 0, "negative_amounts": 0 }, { "day_utc": "2026-09-26", "stored_bills": 22, "hours_with_stored_bills": 20, "recorded_amount_usd": 1.4662088284749188, "unavailable_amounts": 0, "negative_amounts": 0 }, { "day_utc": "2026-09-27", "stored_bills": 23, "hours_with_stored_bills": 20, "recorded_amount_usd": 1.7386752227030229, "unavailable_amounts": 0, "negative_amounts": 0 }, { "day_utc": "2026-09-28", "stored_bills": 23, "hours_with_stored_bills": 23, "recorded_amount_usd": 3.2457063114488847, "unavailable_amounts": 0, "negative_amounts": 0 }, { "day_utc": "2026-09-29", "stored_bills": 23, "hours_with_stored_bills": 23, "recorded_amount_usd": 3.1432410376110056, "unavailable_amounts": 0, "negative_amounts": 0 } ], "currencyVerification": "The billing API documents amount in USD; the stored source field uses USD. No currency conversion is applied.", "freshness": { "latestStoredBillingBucketUtc": "2026-09-30 20:00:00", "latestBillInsertUtc": "2026-09-30 21:01:59", "note": "Insertion timestamps do not update when existing billing records are refreshed." }, "completeness": "Unknown. An hour without stored bills may represent no charges or absent source data; the table has no explicit zero-charge or sync-completeness marker. The source normally refreshes a 48-hour lookback.", "allocation": "No per-job or per-minute cost allocation. These records cover two endpoints and cannot be assigned to a particular settings or duration cohort.", "scope": "Recorded charges only. This is not total operating cost, customer price, a complete cloud invoice or a measured cost per transcription." }, "methodology": { "declaredBeforeTimingReads": true, "declarationTimeUtc": "2026-09-30T21:25:42Z", "dateWindow": { "inclusive": "2026-09-24T00:00:00Z", "exclusive": "2026-09-30T00:00:00Z" }, "timeZone": "UTC", "performanceVersionFilter": { "minimum": "2026-09-23", "format": "whole YYYY-MM-DD release tags", "endpoint": "production inference endpoint, private identifier" }, "settings": { "all_settings": { "definition": "Every recorded settings combination, including custom settings and streaming output." }, "base": { "definition": "Transcription with label audio, parse speakers and parse words off; streaming output and split channel off; confidence 0.5; segment duration 30 seconds; default JSON output.", "labels": false, "speakers": false, "words": false, "streamOutput": false, "splitChannel": false, "minimumConfidence": 0.5, "segmentDurationSeconds": 30, "output": "default JSON" }, "diarization": { "definition": "The same defaults as base, with parse speakers on.", "labels": false, "speakers": true, "words": false, "streamOutput": false, "splitChannel": false, "minimumConfidence": 0.5, "segmentDurationSeconds": 30, "output": "default JSON" } }, "durationBands": { "about_5_minutes": { "minimumSeconds": 270, "maximumSeconds": 330, "minimumMinutes": 4.5, "maximumMinutes": 5.5, "bounds": "inclusive" }, "about_60_minutes": { "minimumSeconds": 3300, "maximumSeconds": 3900, "minimumMinutes": 55, "maximumMinutes": 65, "bounds": "inclusive" } }, "durationBuckets": { "basis": "existing complete Bench duration buckets", "purpose": "broader all-settings context, does not replace the requested bands", "bounds": "lower exclusive, upper inclusive", "seconds": [ { "bucket": "0_to_1_minutes", "lower": 0, "upper": 60 }, { "bucket": "1_to_5_minutes", "lower": 60, "upper": 300 }, { "bucket": "5_to_15_minutes", "lower": 300, "upper": 900 }, { "bucket": "15_to_30_minutes", "lower": 900, "upper": 1800 }, { "bucket": "30_to_60_minutes", "lower": 1800, "upper": 3600 }, { "bucket": "1_to_2_hours", "lower": 3600, "upper": 7200 }, { "bucket": "2_to_4_hours", "lower": 7200, "upper": 14400 }, { "bucket": "4_to_8_hours", "lower": 14400, "upper": 28800 }, { "bucket": "8_to_12_hours", "lower": 28800, "upper": 43200 }, { "bucket": "12_to_24_hours", "lower": 43200, "upper": 86400 }, { "bucket": "24_plus_hours", "lower": 86400, "upper": null } ], "languages": "all languages", "metrics_observed_when_added": false }, "languageScopes": { "all_languages": "No language filter. Detected audio language is unavailable.", "english_requested": "Normalized requested language hint equals en. This does not verify English audio." }, "percentileMethod": "Nearest rank: sort valid observations and select rank ceil(n*p/100). No interpolation. Each metric retains its own sample count, missing count and invalid count.", "validMetrics": "Positive audio duration and job elapsed time; nonnegative inference delay and execution time. Only numeric values qualify.", "failurePolicy": "Show stored-job and known-version outcomes separately from completed-job timing distributions. Never treat missing failure durations as zero failures in a duration band.", "cohortPolicy": "Keep all 18 declared cohort rows and all 11 existing Bench duration buckets, including empty cohorts and all tails. No performance-based exclusions.", "metrics": { "job_elapsed_seconds": { "label": "Server-side job elapsed time", "unit": "seconds", "start": "Job-record creation after URL prefetch and inference dispatch.", "end": "Timestamp captured in the completion handler after updating job status and before persisting analytics, storing the transcript, formatting delivery or sending the customer webhook.", "excludes": [ "Initial request handling and URL prefetch before the job record exists", "Dispatch work before job-record creation", "Transcript storage and customer delivery after the measured timestamp", "Client polling, network receipt and result download" ], "note": "Callback retries can overwrite analytics with a later elapsed measurement. This is not a client submission-to-result timer." }, "provider_delay_seconds": { "label": "Inference delay", "unit": "seconds", "definition": "Delay reported by the inference platform, converted from milliseconds. Includes platform waiting and startup effects. It is not a separate client timer." }, "provider_execution_seconds": { "label": "Inference execution", "unit": "seconds", "definition": "Execution reported by the inference platform, converted from milliseconds. This includes work performed inside the inference job rather than only speech recognition." }, "audio_seconds": { "label": "Recorded audio duration", "unit": "seconds", "definition": "Duration returned by the completed job. Primary presentation should convert seconds to minutes." }, "effectiveSpeed": { "label": "Effective processing speed", "unit": "times audio duration", "definition": "Sum of audio seconds divided by sum of server-side job elapsed seconds for the same completed observations with both values positive. This is a ratio of sums and does not measure concurrent endpoint throughput." } }, "sources": { "benchSourceCommit": "1dacd9b0df57fa76b497fc73c38ed6c8c9d0df93", "benchBoundaryCommit": "918dd3c034d16c799ba8142e2bdbee778f83656f", "jobsSourceCommit": "c944a077257f51f55fdd0310995c721f319d7468", "jobsTimingPaths": [ "src/database.ts", "src/handlers.ts" ], "jobsInputsPath": "src/utils.ts", "biSourceCommit": "2efe03bfd0e587690b9e5eaea9e3175e2cf674a8", "biSyncPath": "worker/schedule.ts", "benchPublicPayload": { "url": "https://bench.speechischeap.com/api/benchmarks", "retrievedOn": "2026-09-30", "generatedAt": "2026-09-30T21:48:22.946Z", "lastModified": "2026-09-30T21:46:14.743Z", "payloadSha256": "8686603bf669e8fbd84f0d022c4407a0de8238241c651a2e63fddeb186cdee6d", "role": "Public Bench aggregate source and version-group conventions. This live payload is not the fixed-window percentile snapshot." } }, "limitations": [ "This is a frozen snapshot of retained production observations, not a controlled same-audio benchmark or a service-level guarantee.", "The elapsed clock begins after URL prefetch and inference dispatch, and ends before transcript storage and customer delivery. Vendor client timers cannot be treated as the same clock.", "Requested English is only a hint. Detected audio language is unavailable. Automatic language requests must not be called English audio.", "The 4.5-5.5 minute English-requested subsets are empty. The 55-65 minute all-settings band has one English-requested observation; its audio language is unverified and its sample is too small for a comparison. All strict base and diarization English-requested subsets are empty. Empty cohorts remain in the data.", "All first and later worker runs are included in main results. No warm-worker filter replaces them.", "The 37 failed latest-group jobs lack duration. They cannot be assigned to duration buckets or used to claim a success rate within a duration band.", "Two completed jobs with older date tags and one failed job with unavailable version metadata remain visible in stored-job coverage and are excluded from latest-group timings.", "The 55-65 minute base cohort has eight observations. Sparse tails do not establish stable population percentiles.", "Completion callbacks upsert analytics. Retained rows can reflect a later callback rather than immutable first-completion history.", "Recorded compute charges do not establish complete or allocated transcription cost." ] } }