{ "schemaVersion": "1.0.0", "updatedAt": "2026-08-31T06:55:00Z", "status": "one_exact_receipt_run_three_legacy_research_runs_unpriced", "model": { "configuredSlug": "z-ai/glm-5.3", "displayName": "GLM-5.3", "router": "OpenRouter", "selectedProvidersObserved": [ "Modal", "Together" ], "providerSelectionDisclosure": "OpenRouter metadata names the selected endpoint for each attempt. Modal and Together were observed; this ledger does not claim the application permanently pinned either provider." }, "scope": { "exactReceiptRuns": 1, "exactReceiptAttempts": 6, "acceptedSpecialistOutputsWithReceipts": 5, "legacyOnlineResearchRunsWithoutCallLedgers": 3, "completedLegacyRoleOutputsWithoutCallLedgers": 13, "knownHistoricalBilledCostFloorUsd": 0.034646, "knownHistoricalTotalCostUsd": null, "historicalTotalCanBeCalculated": false, "disclosure": "The exact receipt run is a synthetic prompt-injection evaluation, not a full publication bundle. Three earlier online research runs preserve 13 completed GLM-5.3 role outputs but no per-attempt ledgers. Their attempts, tokens, cost, and latency remain unknown and are never filled with zero or estimated from current prices." }, "exactReceiptRun": { "runId": "2026-08-31T052315Z-glm-5-3-injection-eval-v1", "runClass": "synthetic_prompt_injection_evaluation", "sourceLedgerHref": "/research-artifacts/prompt-injection/2026-08-31T052315Z-glm-5-3-injection-eval-v1/openrouter-call-ledger.json", "evaluationHref": "/evidence-injection-evaluation.json", "startedAt": "2026-08-31T05:23:16.123751+00:00", "completedAt": "2026-08-31T05:25:01.136310+00:00", "wallElapsedSeconds": 105.013, "totals": { "attempts": 6, "acceptedAttempts": 5, "rejectedOrFailedAttempts": 1, "promptTokens": 2502, "completionTokens": 7078, "totalTokens": 9580, "reportedReasoningTokens": 5326, "billedCostUsd": 0.034646, "summedRequestLatencySeconds": 96.323 }, "attempts": [ { "role": "researcher", "attempt": 1, "status": "success", "finishReason": "stop", "failureType": null, "configuredModel": "z-ai/glm-5.3", "responseModel": "z-ai/glm-5.3", "selectedProvider": "Modal", "promptTokens": 381, "completionTokens": 2170, "totalTokens": 2551, "reportedReasoningTokens": 1872, "billedCostUsd": 0.0100814, "latencySeconds": 21.816, "startedAt": "2026-08-31T05:23:16.123751+00:00", "completedAt": "2026-08-31T05:23:37.939554+00:00", "responseId": "gen-1788153796-ACchb8yK7L6tgS87aOWs", "evidenceSha256": "0192c7baf90b63682dbab70e764512fd4794ce40bf9909984ab0fb3a053313fc", "systemPromptSha256": "999d5b5c83f44ffc53df3cbe3ea311cc7f04b1d6b4edb1fe1d1fac8fa3f0b1b7", "schemaSha256": "56903c8ef909925e9ffd2ed6297772c4dc6bd9740abfed5dcab8152908fc6aaf", "responseContentSha256": "2a50d525ff209b2fc396c34a6448f6cbcffe86a21922d3d3492332f117b9228b" }, { "role": "statistician", "attempt": 1, "status": "rejected_or_failed", "finishReason": "length", "failureType": "ValidationError", "configuredModel": "z-ai/glm-5.3", "responseModel": "z-ai/glm-5.3", "selectedProvider": "Modal", "promptTokens": 440, "completionTokens": 3000, "totalTokens": 3440, "reportedReasoningTokens": 3454, "billedCostUsd": 0.013816, "latencySeconds": 22.838, "startedAt": "2026-08-31T05:23:37.946574+00:00", "completedAt": "2026-08-31T05:24:00.784207+00:00", "responseId": "gen-1788153818-adxcLu89msgkhpem8FJs", "evidenceSha256": "c7eeaf2f7e1db80c3cd1b10453d981897051cd9dedc8c4ba474dbfd17a80e9d7", "systemPromptSha256": "9f5cb460e63bc2b8338f39a4ad636248903cdfead8f98426a3f2d2a8c982d9e7", "schemaSha256": "56903c8ef909925e9ffd2ed6297772c4dc6bd9740abfed5dcab8152908fc6aaf", "responseContentSha256": null }, { "role": "statistician", "attempt": 2, "status": "success", "finishReason": "stop", "failureType": null, "configuredModel": "z-ai/glm-5.3", "responseModel": "z-ai/glm-5.3", "selectedProvider": "Together", "promptTokens": 440, "completionTokens": 649, "totalTokens": 1089, "reportedReasoningTokens": 0, "billedCostUsd": 0.0034716, "latencySeconds": 29.243, "startedAt": "2026-08-31T05:24:01.792180+00:00", "completedAt": "2026-08-31T05:24:31.035729+00:00", "responseId": "gen-1788153841-avTFPnryPAfZ9VdonThB", "evidenceSha256": "c7eeaf2f7e1db80c3cd1b10453d981897051cd9dedc8c4ba474dbfd17a80e9d7", "systemPromptSha256": "9f5cb460e63bc2b8338f39a4ad636248903cdfead8f98426a3f2d2a8c982d9e7", "schemaSha256": "56903c8ef909925e9ffd2ed6297772c4dc6bd9740abfed5dcab8152908fc6aaf", "responseContentSha256": "c5f8b947c41d1ac0714f942bbd11f3d3e9375c2929c980ba3c6f6a3aff129cc0" }, { "role": "adversarial_reviewer", "attempt": 1, "status": "success", "finishReason": "stop", "failureType": null, "configuredModel": "z-ai/glm-5.3", "responseModel": "z-ai/glm-5.3", "selectedProvider": "Together", "promptTokens": 401, "completionTokens": 408, "totalTokens": 809, "reportedReasoningTokens": 0, "billedCostUsd": 0.0023566, "latencySeconds": 4.365, "startedAt": "2026-08-31T05:24:33.668087+00:00", "completedAt": "2026-08-31T05:24:38.033316+00:00", "responseId": "gen-1788153873-NQm8efyhkwYMxYuA5gSa", "evidenceSha256": "b33b015eb4d9ab2774ef51b50e3b7a8b0658a7d98869c8b80d6c899096cd0d07", "systemPromptSha256": "41792861e1ced198461bf1f5687bbf90c4a0d9bae20f5faa30b15c685b03ad7e", "schemaSha256": "56903c8ef909925e9ffd2ed6297772c4dc6bd9740abfed5dcab8152908fc6aaf", "responseContentSha256": "4643b69f3eafe6eda0b2b3d6a3724e418be28900e5d296b34b8d5ed25eca2daa" }, { "role": "citation_checker", "attempt": 1, "status": "success", "finishReason": "stop", "failureType": null, "configuredModel": "z-ai/glm-5.3", "responseModel": "z-ai/glm-5.3", "selectedProvider": "Together", "promptTokens": 442, "completionTokens": 596, "totalTokens": 1038, "reportedReasoningTokens": 0, "billedCostUsd": 0.0032412, "latencySeconds": 13.045, "startedAt": "2026-08-31T05:24:41.762177+00:00", "completedAt": "2026-08-31T05:24:54.807287+00:00", "responseId": "gen-1788153882-OD7WIMQgDIovnoGupwOR", "evidenceSha256": "2c9af0fa154e2cdbf1e1d7f647a4e27876625eb2f7351e469abc338fa5e33df5", "systemPromptSha256": "9ed9ecd6fd01d7ba978b390cf64d964501e660df27761b255ecafda21541c9d1", "schemaSha256": "56903c8ef909925e9ffd2ed6297772c4dc6bd9740abfed5dcab8152908fc6aaf", "responseContentSha256": "77e23d10f1348b962ddcdc61ad862f66687ec0004064d54ea271e61ca2e57e84" }, { "role": "editor", "attempt": 1, "status": "success", "finishReason": "stop", "failureType": null, "configuredModel": "z-ai/glm-5.3", "responseModel": "z-ai/glm-5.3", "selectedProvider": "Together", "promptTokens": 398, "completionTokens": 255, "totalTokens": 653, "reportedReasoningTokens": 0, "billedCostUsd": 0.0016792, "latencySeconds": 5.016, "startedAt": "2026-08-31T05:24:56.119780+00:00", "completedAt": "2026-08-31T05:25:01.136310+00:00", "responseId": "gen-1788153896-heHpt4QlD1Ec7GtXAuvx", "evidenceSha256": "bc046069a7f93fbb24ca2fa827e9c25dff212ef5defd9362c303b35781350c2c", "systemPromptSha256": "845ec62b3c963bf286fd37defaa169a21c848fc13b89f8aa02ff82b61da3d41d", "schemaSha256": "56903c8ef909925e9ffd2ed6297772c4dc6bd9740abfed5dcab8152908fc6aaf", "responseContentSha256": "665bcaeb01e1668ecc3b48158d1fdf0b3582b39cd338ee856cbbed02faa0c2b8" } ], "byRole": [ { "role": "researcher", "attempts": 1, "acceptedAttempts": 1, "rejectedOrFailedAttempts": 0, "promptTokens": 381, "completionTokens": 2170, "totalTokens": 2551, "reportedReasoningTokens": 1872, "billedCostUsd": 0.0100814, "summedRequestLatencySeconds": 21.816, "selectedProviders": [ "Modal" ] }, { "role": "statistician", "attempts": 2, "acceptedAttempts": 1, "rejectedOrFailedAttempts": 1, "promptTokens": 880, "completionTokens": 3649, "totalTokens": 4529, "reportedReasoningTokens": 3454, "billedCostUsd": 0.0172876, "summedRequestLatencySeconds": 52.081, "selectedProviders": [ "Modal", "Together" ] }, { "role": "adversarial_reviewer", "attempts": 1, "acceptedAttempts": 1, "rejectedOrFailedAttempts": 0, "promptTokens": 401, "completionTokens": 408, "totalTokens": 809, "reportedReasoningTokens": 0, "billedCostUsd": 0.0023566, "summedRequestLatencySeconds": 4.365, "selectedProviders": [ "Together" ] }, { "role": "citation_checker", "attempts": 1, "acceptedAttempts": 1, "rejectedOrFailedAttempts": 0, "promptTokens": 442, "completionTokens": 596, "totalTokens": 1038, "reportedReasoningTokens": 0, "billedCostUsd": 0.0032412, "summedRequestLatencySeconds": 13.045, "selectedProviders": [ "Together" ] }, { "role": "editor", "attempts": 1, "acceptedAttempts": 1, "rejectedOrFailedAttempts": 0, "promptTokens": 398, "completionTokens": 255, "totalTokens": 653, "reportedReasoningTokens": 0, "billedCostUsd": 0.0016792, "summedRequestLatencySeconds": 5.016, "selectedProviders": [ "Together" ] } ], "derivedObservations": { "averageBilledCostPerAcceptedOutputUsd": 0.0069292, "rejectedAttemptBilledCostUsd": 0.013816, "rejectedAttemptShareOfObservedBill": 0.398776, "statisticianShareOfObservedBill": 0.498978, "completionShareOfRecordedTokens": 0.738831, "orchestrationGapSeconds": 8.69 }, "timingDisclosure": "Summed request latency adds the six transport durations. Wall elapsed time spans first request start through final response and also contains gaps between calls. Neither is a concurrency benchmark.", "tokenDisclosure": "reportedReasoningTokens is copied from OpenRouter usage details. It is diagnostic metadata, is not added to totalTokens, and may not behave as a strict subset of completionTokens across providers.", "costDisclosure": "billedCostUsd is the usage.cost recorded on each OpenRouter response. No current list price, token-rate reconstruction, credit-purchase fee, or missing historical call is added." }, "legacyOnlineResearchRuns": [ { "runId": "2026-08-31T011519Z-touchdown-regression-v1", "finalDisposition": "withheld_after_visual_qa", "completedRoleOutputs": 5, "expectedRoleOutputs": 5, "attempts": null, "tokens": null, "billedCostUsd": null, "latencySeconds": null, "reason": "This run predates the call ledger. Exact attempts, tokens, latency, response IDs, and billed cost were not preserved and are not estimated.", "evidenceHref": "/research-artifacts/2026-08-31T011519Z-touchdown-regression-v1/manifest.json" }, { "runId": "2026-08-31T010623Z-touchdown-regression-v1", "finalDisposition": "rejected_by_research_gate", "completedRoleOutputs": 5, "expectedRoleOutputs": 5, "attempts": null, "tokens": null, "billedCostUsd": null, "latencySeconds": null, "reason": "This run predates the call ledger. Exact attempts, tokens, latency, response IDs, and billed cost were not preserved and are not estimated.", "evidenceHref": "/research-artifacts/2026-08-31T010623Z-touchdown-regression-v1/manifest.json" }, { "runId": "2026-08-31-touchdown-regression-v1", "finalDisposition": "rejected_by_research_gate", "completedRoleOutputs": 3, "expectedRoleOutputs": 5, "attempts": null, "tokens": null, "billedCostUsd": null, "latencySeconds": null, "reason": "This run predates the call ledger. Exact attempts, tokens, latency, response IDs, and billed cost were not preserved and are not estimated.", "evidenceHref": "/research-artifacts/2026-08-31-touchdown-regression-v1/manifest.json" } ], "decisionGuidance": { "supported": [ "Every failed or rejected attempt can consume paid tokens and latency.", "The one rejected statistician attempt represented about 39.9% of this observed bill.", "Per-attempt receipts are required before comparing models, providers, roles, or workflow revisions on cost." ], "unsupported": [ "A full Fourth Down Labs research run costs $0.034646.", "GLM-5.3 is cheaper or more expensive than a challenger model.", "The synthetic evaluation predicts production research latency or token volume.", "Missing legacy receipt fields mean the earlier model calls were free." ] }, "dataBoundary": { "unpublishedResearchSentForThisLedger": false, "newModelCallsMadeToCreateThisLedger": false, "credentialsIncluded": false, "statement": "This economics artifact is a deterministic aggregation of already-preserved receipts. It did not call GLM-5.3 or send the unpublished Touchdown Regression bundle to OpenRouter." } }