{"provenance":{"authentication":"derived_from_authenticated_public_release","separately_signed":false,"statement":"This sanitized run projection is derived from the authenticated public release and is not separately signed.","urls":{"manifest":"/data/releases/2026.07-calibration/public-manifest.json","manifest_signature":"/data/releases/2026.07-calibration/public-manifest-signature.json","public_keys":[{"key_id":"kaelum-bench-public-release-20260718-v1","url":"/data/keys/kaelum-bench-public-release-20260718-v1"}],"run_index":"/data/releases/2026.07-calibration/public-run-index.json","self":"/data/runs/a24a4d05-27ef-5083-91d0-96148c3a4cfe","source_aggregate":"/data/releases/2026.07-calibration/fmb-live-runs.json"}},"run":{"benchmarkId":"FMB-KNOW-ACC-01","benchmarkVersion":"0.1.0-dev","campaignId":"5d89e353-30a2-5dc4-a347-e529eb3890cb","comparisonGroupId":"accounting-knowledge-0.1.0-dev-configuration-specific-v1","comparisonProtocol":"configuration_specific","completedAt":"2026-07-27T06:41:43.747447Z","configuration":{"comparisonProtocol":"configuration_specific","key":"config-bb2e0f36f468d536259bf55aac3dbccc90ae8d3bae76330120203d122cebc1f6","materialParameters":[{"name":"allow_fallbacks","value":"false"},{"name":"data_collection","value":"deny"},{"name":"max_attempts_per_task","value":"1"},{"name":"max_concurrency","value":"1"},{"name":"max_generation_metadata_lookups","value":"1"},{"name":"max_output_tokens","value":"16384"},{"name":"max_raw_response_bytes","value":"16777216"},{"name":"max_request_bytes","value":"1048576"},{"name":"metadata_timeout_seconds","value":"60"},{"name":"request_timeout_seconds","value":"300"},{"name":"require_parameters","value":"true"},{"name":"temperature","value":"0"},{"name":"zdr","value":"true"}],"providerRoute":"openrouter/deepinfra/base","reasoningEffort":"native-default","reasoningProfile":"native-default","requestedModelId":"meta-llama/llama-4-maverick","runtimeMode":"openrouter-tool-free-non-streaming-v8","toolAccess":"none"},"displayVersion":"v0.2","execution":{"attempted":60,"invalidJson":0,"missing":0,"providerFailed":0,"schemaFailed":0,"tasksTotal":60,"valid":60},"globalReleaseId":"80817257-aa7f-4fb9-867b-66e9ed813de9","judge":{"assessmentNotRequiredTasks":0,"assessmentTasks":60,"costUsd":"1.514568","modelId":"openai/gpt-5.6-sol","providerName":"OpenRouter (Azure EU)","providerRegion":null,"retries":0,"usage":{"inputTokens":112782,"outputTokens":27099,"totalTokens":139881}},"manifest":{"generatedAt":"2026-07-27T07:22:05Z","sha256":"d82a32b09acf027f94de5fe162ca71dabb3a389a4e2b12f58fa3d1faec73a3c8"},"modelDisplayName":"Llama 4 Maverick","providerLabel":"DeepInfra","reasoningEffort":"native-default","releaseSignature":{"algorithm":"ed25519","keyId":"kaelum-bench-public-release-20260718-v1","messageType":"kaelum.bench.fmb.public_release.v1","payloadSha256":"d82a32b09acf027f94de5fe162ca71dabb3a389a4e2b12f58fa3d1faec73a3c8","signedFile":"public-manifest.json"},"requestedModelId":"meta-llama/llama-4-maverick","revisionState":"original","runId":"a24a4d05-27ef-5083-91d0-96148c3a4cfe","score":{"earned":"431.6","maximum":"600","percentage":"71.9333"},"scorerId":"fmb-live-scorer-v1","scoringMethod":"hybrid","selectionState":"primary","signedReleaseId":"0bae13a5-a0c9-44fb-b6df-6bc83a6847d5","sourceAggregate":{"campaignId":"5d89e353-30a2-5dc4-a347-e529eb3890cb","path":"fmb-live-runs.json","sha256":"0ef185b35b7d9f9b891a40de2bc9a9cee0febe77a41ee700458b6b92b4bb2cd3","sizeBytes":17057},"startedAt":"2026-07-27T06:33:04.771021Z","suiteId":"accounting-knowledge","suiteTitle":"Accounting Knowledge","testedModelCostUsd":"0.0127508","toolAccess":"none","totalCostUsd":"1.5273188","usage":{"inputTokens":35510,"outputTokens":7061,"totalTokens":42571},"validity":"provisional"},"schema_version":"kaelum-bench-public-run-projection-v1"}
