{"track": "B2", "stage": "select", "model": "claude-haiku-4-5-20251001", "task": "COMP-05", "rep": 0, "latency_s": 10.63, "prompt_chars": 10869, "n_selected": 1, "n_needed": 1, "tp": 1, "precision": 1.0, "recall": 1.0, "selected": ["meridian.def.net_revenue"], "obj": {"selected": [{"definition_id": "meridian.def.net_revenue", "version": 2}], "note": "Version 2 governs period from 2026-05-01 onward"}, "raw": "```json\n{\"selected\": [{\"definition_id\": \"meridian.def.net_revenue\", \"version\": 2}], \"note\": \"Version 2 governs period from 2026-05-01 onward\"}\n```\n"} {"track": "B2", "stage": "select", "model": "claude-haiku-4-5-20251001", "task": "COMP-07", "rep": 1, "latency_s": 11.57, "prompt_chars": 10849, "n_selected": 1, "n_needed": 1, "tp": 1, "precision": 1.0, "recall": 1.0, "selected": ["meridian.def.delivered"], "obj": {"selected": [{"definition_id": "meridian.def.delivered", "version": 1}], "note": "Internal Meridian reporting; federation records do not apply."}, "raw": "```json\n{\"selected\": [{\"definition_id\": \"meridian.def.delivered\", \"version\": 1}], \"note\": \"Internal Meridian reporting; federation records do not apply.\"}\n```\n"} {"track": "B2", "stage": "select", "model": "claude-haiku-4-5-20251001", "task": "COMP-01", "rep": 0, "latency_s": 12.7, "prompt_chars": 10852, "n_selected": 1, "n_needed": 1, "tp": 1, "precision": 1.0, "recall": 1.0, "selected": ["meridian.def.active_customer"], "obj": {"selected": [{"definition_id": "meridian.def.active_customer", "version": 2}], "note": "Version 2 governs as of 2026-08-31; internal reporting uses Meridian definitions only."}, "raw": "```json\n{\n \"selected\": [\n {\n \"definition_id\": \"meridian.def.active_customer\",\n \"version\": 2\n }\n ],\n \"note\": \"Version 2 governs as of 2026-08-31; internal reporting uses Meridian definitions only.\"\n}\n```\n"} {"track": "B2", "stage": "select", "model": "claude-haiku-4-5-20251001", "task": "COMP-05", "rep": 1, "latency_s": 13.15, "prompt_chars": 10869, "n_selected": 1, "n_needed": 1, "tp": 1, "precision": 1.0, "recall": 1.0, "selected": ["meridian.def.net_revenue"], "obj": {"selected": [{"definition_id": "meridian.def.net_revenue", "version": 2}], "note": "Version 2 effective from 2026-05-01 governs entire period."}, "raw": "```json\n{\"selected\": [{\"definition_id\": \"meridian.def.net_revenue\", \"version\": 2}], \"note\": \"Version 2 effective from 2026-05-01 governs entire period.\"}\n```\n"} {"track": "B2", "stage": "select", "model": "claude-haiku-4-5-20251001", "task": "COMP-07", "rep": 0, "latency_s": 13.37, "prompt_chars": 10849, "n_selected": 1, "n_needed": 1, "tp": 1, "precision": 1.0, "recall": 1.0, "selected": ["meridian.def.delivered"], "obj": {"selected": [{"definition_id": "meridian.def.delivered", "version": 1}], "note": "Internal Meridian reporting applies meridian definitions; delivered v1 in force."}, "raw": "```json\n{\n \"selected\": [\n {\n \"definition_id\": \"meridian.def.delivered\",\n \"version\": 1\n }\n ],\n \"note\": \"Internal Meridian reporting applies meridian definitions; delivered v1 in force.\"\n}\n```\n"} {"track": "B2", "stage": "select", "model": "claude-haiku-4-5-20251001", "task": "COMP-10", "rep": 1, "latency_s": 14.02, "prompt_chars": 10857, "n_selected": 2, "n_needed": 1, "tp": 1, "precision": 0.5, "recall": 1.0, "selected": ["meridian.def.delivered", "meridian.def.on_time"], "obj": {"selected": [{"definition_id": "meridian.def.delivered", "version": 1}, {"definition_id": "meridian.def.on_time", "version": 2}], "note": "Internal Meridian reporting as-of 2026-08-31: definitions of delivered and on-time"}, "raw": "```json\n{\n \"selected\": [\n {\"definition_id\": \"meridian.def.delivered\", \"version\": 1},\n {\"definition_id\": \"meridian.def.on_time\", \"version\": 2}\n ],\n \"note\": \"Internal Meridian reporting as-of 2026-08-31: definitions of delivered and on-time\"\n}\n```\n"} {"track": "B2", "stage": "select", "model": "claude-haiku-4-5-20251001", "task": "COMP-10", "rep": 0, "latency_s": 14.61, "prompt_chars": 10857, "n_selected": 2, "n_needed": 1, "tp": 1, "precision": 0.5, "recall": 1.0, "selected": ["meridian.def.delivered", "meridian.def.on_time"], "obj": {"selected": [{"definition_id": "meridian.def.delivered", "version": 1}, {"definition_id": "meridian.def.on_time", "version": 2}], "note": "Delivered and on-time for internal Meridian reporting as of 2026-08-31."}, "raw": "```json\n{\n \"selected\": [\n {\"definition_id\": \"meridian.def.delivered\", \"version\": 1},\n {\"definition_id\": \"meridian.def.on_time\", \"version\": 2}\n ],\n \"note\": \"Delivered and on-time for internal Meridian reporting as of 2026-08-31.\"\n}\n```\n"} {"track": "B2", "stage": "select", "model": "claude-haiku-4-5-20251001", "task": "COMP-01", "rep": 1, "latency_s": 16.91, "prompt_chars": 10852, "n_selected": 1, "n_needed": 1, "tp": 1, "precision": 1.0, "recall": 1.0, "selected": ["meridian.def.active_customer"], "obj": {"selected": [{"definition_id": "meridian.def.active_customer", "version": 2}], "note": "Version 2 applies on 2026-08-31; window_days parameter is 180"}, "raw": "```json\n{\n \"selected\": [\n {\n \"definition_id\": \"meridian.def.active_customer\",\n \"version\": 2\n }\n ],\n \"note\": \"Version 2 applies on 2026-08-31; window_days parameter is 180\"\n}\n```\n"} {"track": "B2", "stage": "select", "model": "claude-haiku-4-5-20251001", "task": "TEMP-01", "rep": 1, "latency_s": 10.73, "prompt_chars": 10852, "n_selected": 1, "n_needed": 1, "tp": 1, "precision": 1.0, "recall": 1.0, "selected": ["meridian.def.active_customer"], "obj": {"selected": [{"definition_id": "meridian.def.active_customer", "version": 1}], "note": "Version 1 effective on 2026-02-28; version 2 begins 2026-04-01"}, "raw": "```json\n{\n \"selected\": [\n {\n \"definition_id\": \"meridian.def.active_customer\",\n \"version\": 1\n }\n ],\n \"note\": \"Version 1 effective on 2026-02-28; version 2 begins 2026-04-01\"\n}\n```\n"} {"track": "B2", "stage": "select", "model": "claude-haiku-4-5-20251001", "task": "COMP-12", "rep": 1, "latency_s": 15.01, "prompt_chars": 10903, "n_selected": 1, "n_needed": 1, "tp": 1, "precision": 1.0, "recall": 1.0, "selected": ["meridian.def.order_value"], "obj": {"selected": [{"definition_id": "meridian.def.order_value", "version": 1}], "note": "Order value definition for internal Meridian reporting"}, "raw": "```json\n{\n \"selected\": [\n {\n \"definition_id\": \"meridian.def.order_value\",\n \"version\": 1\n }\n ],\n \"note\": \"Order value definition for internal Meridian reporting\"\n}\n```\n"} {"track": "B2", "stage": "select", "model": "claude-haiku-4-5-20251001", "task": "TEMP-07", "rep": 1, "latency_s": 13.86, "prompt_chars": 10869, "n_selected": 1, "n_needed": 1, "tp": 1, "precision": 1.0, "recall": 1.0, "selected": ["meridian.def.net_revenue"], "obj": {"selected": [{"definition_id": "meridian.def.net_revenue", "version": 1}], "note": "Version 1 effective through 2026-04-30 covers entire question period"}, "raw": "```json\n{\n \"selected\": [\n {\n \"definition_id\": \"meridian.def.net_revenue\",\n \"version\": 1\n }\n ],\n \"note\": \"Version 1 effective through 2026-04-30 covers entire question period\"\n}\n```\n"} {"track": "B2", "stage": "select", "model": "claude-haiku-4-5-20251001", "task": "COMP-12", "rep": 0, "latency_s": 19.83, "prompt_chars": 10903, "n_selected": 1, "n_needed": 1, "tp": 1, "precision": 1.0, "recall": 1.0, "selected": ["meridian.def.order_value"], "obj": {"selected": [{"definition_id": "meridian.def.order_value", "version": 2}], "note": "Order value v2 in force as of 2026-07-31"}, "raw": "```json\n{\n \"selected\": [{\"definition_id\": \"meridian.def.order_value\", \"version\": 2}],\n \"note\": \"Order value v2 in force as of 2026-07-31\"\n}\n```\n"} {"track": "B2", "stage": "select", "model": "claude-haiku-4-5-20251001", "task": "TEMP-01", "rep": 0, "latency_s": 19.09, "prompt_chars": 10852, "n_selected": 1, "n_needed": 1, "tp": 1, "precision": 1.0, "recall": 1.0, "selected": ["meridian.def.active_customer"], "obj": {"selected": [{"definition_id": "meridian.def.active_customer", "version": 1}], "note": "Version 1 applies; active_customer is the only definition required"}, "raw": "```json\n{\n \"selected\": [\n {\n \"definition_id\": \"meridian.def.active_customer\",\n \"version\": 1\n }\n ],\n \"note\": \"Version 1 applies; active_customer is the only definition required\"\n}\n```\n"} {"track": "B2", "stage": "select", "model": "claude-haiku-4-5-20251001", "task": "XORG-01", "rep": 0, "latency_s": 19.42, "prompt_chars": 10897, "n_selected": 1, "n_needed": 1, "tp": 1, "precision": 1.0, "recall": 1.0, "selected": ["halden.def.delivered"], "obj": {"selected": [{"definition_id": "halden.def.delivered", "version": 2}], "note": "Halden delivered definition governs C-118 reconciliation"}, "raw": "```json\n{\n \"selected\": [{\"definition_id\": \"halden.def.delivered\", \"version\": 2}],\n \"note\": \"Halden delivered definition governs C-118 reconciliation\"\n}\n```\n"} {"track": "B2", "stage": "select", "model": "claude-haiku-4-5-20251001", "task": "TEMP-07", "rep": 0, "latency_s": 25.89, "prompt_chars": 10869, "n_selected": 1, "n_needed": 1, "tp": 1, "precision": 1.0, "recall": 1.0, "selected": ["meridian.def.net_revenue"], "obj": {"selected": [{"definition_id": "meridian.def.net_revenue", "version": 1}], "note": "Version 1 effective through 2026-04-30; internal Meridian only"}, "raw": "```json\n{\"selected\": [{\"definition_id\": \"meridian.def.net_revenue\", \"version\": 1}], \"note\": \"Version 1 effective through 2026-04-30; internal Meridian only\"}\n```\n"} {"track": "B2", "stage": "select", "model": "claude-haiku-4-5-20251001", "task": "XORG-04", "rep": 0, "latency_s": 16.38, "prompt_chars": 10905, "n_selected": 2, "n_needed": 2, "tp": 2, "precision": 1.0, "recall": 1.0, "selected": ["halden.def.delivered", "meridian.def.on_time"], "obj": {"selected": [{"definition_id": "halden.def.delivered", "version": 2}, {"definition_id": "meridian.def.on_time", "version": 2}], "note": "Delivered per C-118, on-time per Meridian as of 2026-08-31"}, "raw": "```json\n{\n \"selected\": [\n {\"definition_id\": \"halden.def.delivered\", \"version\": 2},\n {\"definition_id\": \"meridian.def.on_time\", \"version\": 2}\n ],\n \"note\": \"Delivered per C-118, on-time per Meridian as of 2026-08-31\"\n}\n```\n"} {"track": "B2", "stage": "select", "model": "claude-haiku-4-5-20251001", "task": "XORG-04", "rep": 1, "latency_s": 14.13, "prompt_chars": 10905, "n_selected": 2, "n_needed": 2, "tp": 2, "precision": 1.0, "recall": 1.0, "selected": ["halden.def.delivered", "meridian.def.on_time"], "obj": {"selected": [{"definition_id": "halden.def.delivered", "version": 2}, {"definition_id": "meridian.def.on_time", "version": 2}], "note": "Halden-governed delivered; meridian on-time for C-118 as of 2026-08-31"}, "raw": "```json\n{\n \"selected\": [\n {\"definition_id\": \"halden.def.delivered\", \"version\": 2},\n {\"definition_id\": \"meridian.def.on_time\", \"version\": 2}\n ],\n \"note\": \"Halden-governed delivered; meridian on-time for C-118 as of 2026-08-31\"\n}\n```\n"} {"track": "B2", "stage": "select", "model": "claude-haiku-4-5-20251001", "task": "XORG-09", "rep": 0, "latency_s": 32.59, "prompt_chars": 11030, "n_selected": 1, "n_needed": 1, "tp": 1, "precision": 1.0, "recall": 1.0, "selected": ["halden.def.business_day"], "obj": {"selected": [{"definition_id": "halden.def.business_day", "version": 1}], "note": "Defines business days for Halden C-118 SLA"}, "raw": "```json\n{\"selected\": [{\"definition_id\": \"halden.def.business_day\", \"version\": 1}], \"note\": \"Defines business days for Halden C-118 SLA\"}\n```\n"} {"track": "B2", "stage": "select", "model": "claude-haiku-4-5-20251001", "task": "XORG-09", "rep": 1, "latency_s": 33.67, "prompt_chars": 11030, "n_selected": 1, "n_needed": 1, "tp": 1, "precision": 1.0, "recall": 1.0, "selected": ["halden.def.business_day"], "obj": {"selected": [{"definition_id": "halden.def.business_day", "version": 1}], "note": "Defines business days for Halden SLA with 2026 holiday schedule."}, "raw": "```json\n{\n \"selected\": [\n {\n \"definition_id\": \"halden.def.business_day\",\n \"version\": 1\n }\n ],\n \"note\": \"Defines business days for Halden SLA with 2026 holiday schedule.\"\n}\n```\n"} {"track": "B2", "stage": "select", "model": "claude-haiku-4-5-20251001", "task": "XORG-01", "rep": 1, "latency_s": 57.86, "prompt_chars": 10897, "n_selected": 1, "n_needed": 1, "tp": 1, "precision": 1.0, "recall": 1.0, "selected": ["halden.def.delivered"], "obj": {"selected": [{"definition_id": "halden.def.delivered", "version": 2}], "note": "Halden definition governs C-118 reconciliation."}, "raw": "```json\n{\n \"selected\": [\n {\"definition_id\": \"halden.def.delivered\", \"version\": 2}\n ],\n \"note\": \"Halden definition governs C-118 reconciliation.\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-05", "cond": "self_selected", "rep": 1, "latency_s": 32.29, "prompt_chars": 11588, "context_chars": 500, "answer_correct": 0, "obj": {"answer": 91300, "definition_id": "meridian.def.net_revenue", "definition_version": 2, "unanswerable": false, "note": "Invoices dated 2026-05-01 to 2026-06-30: sum of (gross - discount + freight) = 97850; returns in period = 6550"}, "raw": "```json\n{\n \"answer\": 91300,\n \"definition_id\": \"meridian.def.net_revenue\",\n \"definition_version\": 2,\n \"unanswerable\": false,\n \"note\": \"Invoices dated 2026-05-01 to 2026-06-30: sum of (gross - discount + freight) = 97850; returns in period = 6550\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "XORG-09", "cond": "full_catalogue", "rep": 0, "latency_s": 32.51, "prompt_chars": 21581, "context_chars": 10332, "answer_correct": 1, "obj": {"answer": 9, "definition_id": "halden.def.business_day", "definition_version": 1, "unanswerable": false, "note": "O006: 2025-12-30 to 2026-01-14. Business days excluding weekends and Halden holidays (2026-01-01, 2026-01-06)."}, "raw": "```json\n{\n \"answer\": 9,\n \"definition_id\": \"halden.def.business_day\",\n \"definition_version\": 1,\n \"unanswerable\": false,\n \"note\": \"O006: 2025-12-30 to 2026-01-14. Business days excluding weekends and Halden holidays (2026-01-01, 2026-01-06).\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "TEMP-07", "cond": "full_catalogue", "rep": 1, "latency_s": 32.67, "prompt_chars": 21420, "context_chars": 10332, "answer_correct": 1, "obj": {"answer": 109343, "definition_id": "meridian.def.net_revenue", "definition_version": 1, "unanswerable": false, "note": "Q1 2026 invoiced less returns, v1 excl freight."}, "raw": "```json\n{\n \"answer\": 109343,\n \"definition_id\": \"meridian.def.net_revenue\",\n \"definition_version\": 1,\n \"unanswerable\": false,\n \"note\": \"Q1 2026 invoiced less returns, v1 excl freight.\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "XORG-01", "cond": "full_catalogue", "rep": 1, "latency_s": 32.87, "prompt_chars": 21448, "context_chars": 10332, "answer_correct": 0, "obj": {"answer": 17, "definition_id": "halden.def.delivered", "definition_version": 2, "unanswerable": false, "note": "Counted C-118 shipments with carrier_scan_date present and on/before cutoff date"}, "raw": "```json\n{\n \"answer\": 17,\n \"definition_id\": \"halden.def.delivered\",\n \"definition_version\": 2,\n \"unanswerable\": false,\n \"note\": \"Counted C-118 shipments with carrier_scan_date present and on/before cutoff date\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "XORG-09", "cond": "full_catalogue", "rep": 1, "latency_s": 41.58, "prompt_chars": 21581, "context_chars": 10332, "answer_correct": 1, "obj": {"answer": 9, "definition_id": "halden.def.business_day", "definition_version": 1, "unanswerable": false, "note": "Halden business days excluding holidays and weekends"}, "raw": "```json\n{\n \"answer\": 9,\n \"definition_id\": \"halden.def.business_day\",\n \"definition_version\": 1,\n \"unanswerable\": false,\n \"note\": \"Halden business days excluding holidays and weekends\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-05", "cond": "self_selected", "rep": 0, "latency_s": 56.16, "prompt_chars": 11588, "context_chars": 500, "answer_correct": 1, "obj": {"answer": 104670, "definition_id": "meridian.def.net_revenue", "definition_version": 2, "unanswerable": false, "note": "Period invoices minus returns for 2026-05-01 to 2026-06-30"}, "raw": "```json\n{\n \"answer\": 104670,\n \"definition_id\": \"meridian.def.net_revenue\",\n \"definition_version\": 2,\n \"unanswerable\": false,\n \"note\": \"Period invoices minus returns for 2026-05-01 to 2026-06-30\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-05", "cond": "keyword_top5", "rep": 1, "latency_s": 26.19, "prompt_chars": 13158, "context_chars": 2070, "answer_correct": 1, "obj": {"answer": 104670, "definition_id": "meridian.def.net_revenue", "definition_version": 2, "unanswerable": false, "note": "Invoices dated 2026-05-01 to 2026-06-30, minus returns in same period."}, "raw": "```json\n{\n \"answer\": 104670,\n \"definition_id\": \"meridian.def.net_revenue\",\n \"definition_version\": 2,\n \"unanswerable\": false,\n \"note\": \"Invoices dated 2026-05-01 to 2026-06-30, minus returns in same period.\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-12", "cond": "full_catalogue", "rep": 1, "latency_s": 31.73, "prompt_chars": 21454, "context_chars": 10332, "answer_correct": 1, "obj": {"answer": 139070, "definition_id": "meridian.def.order_value", "definition_version": 1, "unanswerable": false, "note": "Sum of gross_amount minus discount_amount for 15 non-cancelled orders"}, "raw": "```json\n{\"answer\": 139070, \"definition_id\": \"meridian.def.order_value\", \"definition_version\": 1, \"unanswerable\": false, \"note\": \"Sum of gross_amount minus discount_amount for 15 non-cancelled orders\"}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "TEMP-01", "cond": "keyword_top5", "rep": 1, "latency_s": 67.95, "prompt_chars": 13249, "context_chars": 2178, "answer_correct": 1, "obj": {"answer": 21, "definition_id": "meridian.def.active_customer", "definition_version": 1, "unanswerable": false, "note": "Version 1 applies (effective until 2026-03-31); window_days=365"}, "raw": "```json\n{\n \"answer\": 21,\n \"definition_id\": \"meridian.def.active_customer\",\n \"definition_version\": 1,\n \"unanswerable\": false,\n \"note\": \"Version 1 applies (effective until 2026-03-31); window_days=365\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "XORG-01", "cond": "self_selected", "rep": 1, "latency_s": 29.26, "prompt_chars": 11469, "context_chars": 353, "answer_correct": 1, "obj": {"answer": 18, "definition_id": "halden.def.delivered", "definition_version": 2, "unanswerable": false, "note": "C-118 shipments with carrier_scan_date on/before 2026-08-31"}, "raw": "```json\n{\n \"answer\": 18,\n \"definition_id\": \"halden.def.delivered\",\n \"definition_version\": 2,\n \"unanswerable\": false,\n \"note\": \"C-118 shipments with carrier_scan_date on/before 2026-08-31\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "XORG-04", "cond": "full_catalogue", "rep": 1, "latency_s": 90.43, "prompt_chars": 21456, "context_chars": 10332, "answer_correct": 0, "obj": {"answer": 12, "definition_id": "meridian.def.on_time", "definition_version": 2, "unanswerable": false, "note": "C-118 shipments delivered as of 2026-08-31 meeting on-time criteria under Halden federation rules."}, "raw": "```json\n{\n \"answer\": 12,\n \"definition_id\": \"meridian.def.on_time\",\n \"definition_version\": 2,\n \"unanswerable\": false,\n \"note\": \"C-118 shipments delivered as of 2026-08-31 meeting on-time criteria under Halden federation rules.\"\n}\n```\n\n**Methodology:**\n\n1. **Identified C-118 orders with Halden (HAL)**: 18 orders with shipments.\n\n2. **Applied delivery definition** (fed.C-118.delivered → halden.d"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "TEMP-01", "cond": "self_selected", "rep": 0, "latency_s": 38.84, "prompt_chars": 11562, "context_chars": 491, "answer_correct": 1, "obj": {"answer": 21, "definition_id": "meridian.def.active_customer", "definition_version": 1, "unanswerable": false, "note": "21 customers had non-cancelled orders within 365 days before 2026-02-28."}, "raw": "```json\n{\n \"answer\": 21,\n \"definition_id\": \"meridian.def.active_customer\",\n \"definition_version\": 1,\n \"unanswerable\": false,\n \"note\": \"21 customers had non-cancelled orders within 365 days before 2026-02-28.\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "TEMP-07", "cond": "self_selected", "rep": 0, "latency_s": 32.86, "prompt_chars": 11593, "context_chars": 505, "answer_correct": 0, "obj": {"answer": 108623, "definition_id": "meridian.def.net_revenue", "definition_version": 1, "unanswerable": false, "note": "Q1 2026 invoices less returns per definition"}, "raw": "```json\n{\n \"answer\": 108623,\n \"definition_id\": \"meridian.def.net_revenue\",\n \"definition_version\": 1,\n \"unanswerable\": false,\n \"note\": \"Q1 2026 invoices less returns per definition\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-10", "cond": "self_selected", "rep": 0, "latency_s": 65.39, "prompt_chars": 11750, "context_chars": 674, "answer_correct": 1, "obj": {"answer": 23, "definition_id": "meridian.def.on_time", "definition_version": 2, "unanswerable": false, "note": "23 of 50 delivered shipments met on-time criteria."}, "raw": "```json\n{\n \"answer\": 23,\n \"definition_id\": \"meridian.def.on_time\",\n \"definition_version\": 2,\n \"unanswerable\": false,\n \"note\": \"23 of 50 delivered shipments met on-time criteria.\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "TEMP-01", "cond": "full_catalogue", "rep": 1, "latency_s": 85.23, "prompt_chars": 21403, "context_chars": 10332, "answer_correct": 1, "obj": {"answer": 21, "definition_id": "meridian.def.active_customer", "definition_version": 1, "unanswerable": false, "note": "Version 1 applies; window_days=365; interval (2025-02-28, 2026-02-28]"}, "raw": "```json\n{\n \"answer\": 21,\n \"definition_id\": \"meridian.def.active_customer\",\n \"definition_version\": 1,\n \"unanswerable\": false,\n \"note\": \"Version 1 applies; window_days=365; interval (2025-02-28, 2026-02-28]\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-07", "cond": "self_selected", "rep": 0, "latency_s": 22.99, "prompt_chars": 11419, "context_chars": 351, "answer_correct": 0, "obj": {"answer": 51, "definition_id": "meridian.def.delivered", "definition_version": 1, "unanswerable": false, "note": "51 shipments have customer_confirm_date on/before 2026-08-31"}, "raw": "```json\n{\"answer\": 51, \"definition_id\": \"meridian.def.delivered\", \"definition_version\": 1, \"unanswerable\": false, \"note\": \"51 shipments have customer_confirm_date on/before 2026-08-31\"}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "XORG-09", "cond": "self_selected", "rep": 0, "latency_s": 27.85, "prompt_chars": 11706, "context_chars": 457, "answer_correct": 1, "obj": {"answer": 9, "definition_id": "halden.def.business_day", "definition_version": 1, "unanswerable": false, "note": "Counted weekdays excluding holidays between order date and scan date."}, "raw": "```json\n{\"answer\": 9, \"definition_id\": \"halden.def.business_day\", \"definition_version\": 1, \"unanswerable\": false, \"note\": \"Counted weekdays excluding holidays between order date and scan date.\"}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "XORG-09", "cond": "keyword_top5", "rep": 0, "latency_s": 61.25, "prompt_chars": 12730, "context_chars": 1481, "answer_correct": 0, "obj": {"answer": 10, "definition_id": null, "definition_version": null, "unanswerable": false, "note": "Mon-Fri only; 2026-01-01 treated as holiday"}, "raw": "```json\n{\n \"answer\": 10,\n \"definition_id\": null,\n \"definition_version\": null,\n \"unanswerable\": false,\n \"note\": \"Mon-Fri only; 2026-01-01 treated as holiday\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-10", "cond": "full_catalogue", "rep": 0, "latency_s": 43.2, "prompt_chars": 21408, "context_chars": 10332, "answer_correct": 0, "obj": {"answer": 24, "definition_id": "meridian.def.on_time", "definition_version": 2, "unanswerable": false, "note": "24 delivered shipments met on-time criteria as of 2026-08-31"}, "raw": "```json\n{\"answer\": 24, \"definition_id\": \"meridian.def.on_time\", \"definition_version\": 2, \"unanswerable\": false, \"note\": \"24 delivered shipments met on-time criteria as of 2026-08-31\"}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "XORG-01", "cond": "full_catalogue", "rep": 0, "latency_s": 110.11, "prompt_chars": 21448, "context_chars": 10332, "answer_correct": 1, "obj": {"answer": 18, "definition_id": "halden.def.delivered", "definition_version": 2, "unanswerable": false, "note": "C-118 shipments with carrier_scan_date on/before 2026-08-31"}, "raw": "```json\n{\n \"answer\": 18,\n \"definition_id\": \"halden.def.delivered\",\n \"definition_version\": 2,\n \"unanswerable\": false,\n \"note\": \"C-118 shipments with carrier_scan_date on/before 2026-08-31\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-12", "cond": "keyword_top5", "rep": 0, "latency_s": 35.03, "prompt_chars": 13090, "context_chars": 1968, "answer_correct": 0, "obj": {"answer": 126340, "definition_id": "meridian.def.order_value", "definition_version": 1, "unanswerable": false, "note": "14 orders placed 2026-05-01 to 2026-07-31, all non-cancelled"}, "raw": "```json\n{\n \"answer\": 126340,\n \"definition_id\": \"meridian.def.order_value\",\n \"definition_version\": 1,\n \"unanswerable\": false,\n \"note\": \"14 orders placed 2026-05-01 to 2026-07-31, all non-cancelled\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-01", "cond": "self_selected", "rep": 0, "latency_s": 44.93, "prompt_chars": 11559, "context_chars": 488, "answer_correct": 0, "obj": {"answer": 17, "definition_id": "meridian.def.active_customer", "definition_version": 2, "unanswerable": false, "note": "Orders placed 2026-03-05 to 2026-08-31, status not cancelled."}, "raw": "```json\n{\"answer\": 17, \"definition_id\": \"meridian.def.active_customer\", \"definition_version\": 2, \"unanswerable\": false, \"note\": \"Orders placed 2026-03-05 to 2026-08-31, status not cancelled.\"}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-10", "cond": "full_catalogue", "rep": 1, "latency_s": 56.84, "prompt_chars": 21408, "context_chars": 10332, "answer_correct": 1, "obj": {"answer": 23, "definition_id": "meridian.def.on_time", "definition_version": 2, "unanswerable": false, "note": "23 shipments with confirmed delivery and on-time performance as of 2026-08-31"}, "raw": "```json\n{\n \"answer\": 23,\n \"definition_id\": \"meridian.def.on_time\",\n \"definition_version\": 2,\n \"unanswerable\": false,\n \"note\": \"23 shipments with confirmed delivery and on-time performance as of 2026-08-31\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "XORG-01", "cond": "keyword_top5", "rep": 0, "latency_s": 40.53, "prompt_chars": 13108, "context_chars": 1992, "answer_correct": 0, "obj": {"answer": 19, "definition_id": "halden.def.delivered", "definition_version": 2, "unanswerable": false, "note": "Counted C-118 shipments with carrier_scan_date present and on/before 2026-08-31"}, "raw": "```json\n{\n \"answer\": 19,\n \"definition_id\": \"halden.def.delivered\",\n \"definition_version\": 2,\n \"unanswerable\": false,\n \"note\": \"Counted C-118 shipments with carrier_scan_date present and on/before 2026-08-31\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-07", "cond": "full_catalogue", "rep": 0, "latency_s": 35.77, "prompt_chars": 21400, "context_chars": 10332, "answer_correct": 1, "obj": {"answer": 50, "definition_id": "meridian.def.delivered", "definition_version": 1, "unanswerable": false, "note": "Counted shipments with customer_confirm_date present and on/before 2026-08-31"}, "raw": "```json\n{\"answer\": 50, \"definition_id\": \"meridian.def.delivered\", \"definition_version\": 1, \"unanswerable\": false, \"note\": \"Counted shipments with customer_confirm_date present and on/before 2026-08-31\"}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "XORG-09", "cond": "keyword_top5", "rep": 1, "latency_s": 27.53, "prompt_chars": 12730, "context_chars": 1481, "answer_correct": 0, "obj": {"answer": null, "definition_id": "fed.C-118.business_day", "definition_version": 1, "unanswerable": true, "note": "Governing definition halden.def.business_day not provided in data"}, "raw": "```json\n{\n \"answer\": null,\n \"definition_id\": \"fed.C-118.business_day\",\n \"definition_version\": 1,\n \"unanswerable\": true,\n \"note\": \"Governing definition halden.def.business_day not provided in data\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-01", "cond": "keyword_top5", "rep": 0, "latency_s": 57.27, "prompt_chars": 13249, "context_chars": 2178, "answer_correct": 1, "obj": {"answer": 18, "definition_id": "meridian.def.active_customer", "definition_version": 2, "unanswerable": false, "note": "180-day window from 2026-03-04 to 2026-08-31, non-cancelled orders only"}, "raw": "```json\n{\n \"answer\": 18,\n \"definition_id\": \"meridian.def.active_customer\",\n \"definition_version\": 2,\n \"unanswerable\": false,\n \"note\": \"180-day window from 2026-03-04 to 2026-08-31, non-cancelled orders only\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-10", "cond": "self_selected", "rep": 1, "latency_s": 79.47, "prompt_chars": 11750, "context_chars": 674, "answer_correct": 1, "obj": {"answer": 23, "definition_id": "meridian.def.on_time", "definition_version": 2, "unanswerable": false, "note": "23 of 50 delivered shipments met promised/revised-promised dates."}, "raw": "```json\n{\n \"answer\": 23,\n \"definition_id\": \"meridian.def.on_time\",\n \"definition_version\": 2,\n \"unanswerable\": false,\n \"note\": \"23 of 50 delivered shipments met promised/revised-promised dates.\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-07", "cond": "keyword_top5", "rep": 1, "latency_s": 25.29, "prompt_chars": 13003, "context_chars": 1935, "answer_correct": 1, "obj": {"answer": 50, "definition_id": "meridian.def.delivered", "definition_version": 1, "unanswerable": false, "note": "Counted shipments where customer_confirm_date present and ≤2026-08-31"}, "raw": "```json\n{\n \"answer\": 50,\n \"definition_id\": \"meridian.def.delivered\",\n \"definition_version\": 1,\n \"unanswerable\": false,\n \"note\": \"Counted shipments where customer_confirm_date present and ≤2026-08-31\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-12", "cond": "self_selected", "rep": 0, "latency_s": 27.35, "prompt_chars": 11453, "context_chars": 331, "answer_correct": 1, "obj": {"answer": 139070, "definition_id": "meridian.def.order_value", "definition_version": 1, "unanswerable": false, "note": "15 non-cancelled orders placed in period, summed (gross - discount)."}, "raw": "```json\n{\n \"answer\": 139070,\n \"definition_id\": \"meridian.def.order_value\",\n \"definition_version\": 1,\n \"unanswerable\": false,\n \"note\": \"15 non-cancelled orders placed in period, summed (gross - discount).\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-07", "cond": "full_catalogue", "rep": 1, "latency_s": 28.79, "prompt_chars": 21400, "context_chars": 10332, "answer_correct": 1, "obj": {"answer": 50, "definition_id": "meridian.def.delivered", "definition_version": 1, "unanswerable": false, "note": "Shipments with customer_confirm_date on/before 2026-08-31"}, "raw": "```json\n{\"answer\": 50, \"definition_id\": \"meridian.def.delivered\", \"definition_version\": 1, \"unanswerable\": false, \"note\": \"Shipments with customer_confirm_date on/before 2026-08-31\"}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-12", "cond": "self_selected", "rep": 1, "latency_s": 46.17, "prompt_chars": 11253, "context_chars": 131, "answer_correct": 0, "obj": {"answer": 153700, "definition_id": null, "definition_version": null, "unanswerable": false, "note": "15 orders placed 2026-05-01 to 2026-07-31, all non-cancelled"}, "raw": "```json\n{\n \"answer\": 153700,\n \"definition_id\": null,\n \"definition_version\": null,\n \"unanswerable\": false,\n \"note\": \"15 orders placed 2026-05-01 to 2026-07-31, all non-cancelled\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-01", "cond": "keyword_top5", "rep": 1, "latency_s": 53.53, "prompt_chars": 13249, "context_chars": 2178, "answer_correct": 1, "obj": {"answer": 18, "definition_id": "meridian.def.active_customer", "definition_version": 2, "unanswerable": false, "note": "180-day window (2026-03-05 to 2026-08-31), non-cancelled orders only"}, "raw": "```json\n{\n \"answer\": 18,\n \"definition_id\": \"meridian.def.active_customer\",\n \"definition_version\": 2,\n \"unanswerable\": false,\n \"note\": \"180-day window (2026-03-05 to 2026-08-31), non-cancelled orders only\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "TEMP-07", "cond": "keyword_top5", "rep": 0, "latency_s": 39.35, "prompt_chars": 13158, "context_chars": 2070, "answer_correct": 1, "obj": {"answer": 109343, "definition_id": "meridian.def.net_revenue", "definition_version": 1, "unanswerable": false, "note": "Q1 2026 invoices (13 total) sum (gross-discount); one return R002 deducted."}, "raw": "```json\n{\"answer\": 109343, \"definition_id\": \"meridian.def.net_revenue\", \"definition_version\": 1, \"unanswerable\": false, \"note\": \"Q1 2026 invoices (13 total) sum (gross-discount); one return R002 deducted.\"}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "XORG-04", "cond": "full_catalogue", "rep": 0, "latency_s": 57.85, "prompt_chars": 21456, "context_chars": 10332, "answer_correct": 0, "obj": {"answer": 8, "definition_id": "meridian.def.on_time", "definition_version": 2, "unanswerable": false, "note": "13 C-118 HAL shipments delivered; 8 on-time per revised_promised_date"}, "raw": "```json\n{\n \"answer\": 8,\n \"definition_id\": \"meridian.def.on_time\",\n \"definition_version\": 2,\n \"unanswerable\": false,\n \"note\": \"13 C-118 HAL shipments delivered; 8 on-time per revised_promised_date\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-12", "cond": "full_catalogue", "rep": 0, "latency_s": 24.32, "prompt_chars": 21454, "context_chars": 10332, "answer_correct": 1, "obj": {"answer": 139070, "definition_id": "meridian.def.order_value", "definition_version": 1, "unanswerable": false, "note": "15 non-cancelled orders between 2026-05-01 and 2026-07-31 inclusive"}, "raw": "```json\n{\n \"answer\": 139070,\n \"definition_id\": \"meridian.def.order_value\",\n \"definition_version\": 1,\n \"unanswerable\": false,\n \"note\": \"15 non-cancelled orders between 2026-05-01 and 2026-07-31 inclusive\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "TEMP-07", "cond": "full_catalogue", "rep": 0, "latency_s": 40.88, "prompt_chars": 21420, "context_chars": 10332, "answer_correct": 1, "obj": {"answer": 109343, "definition_id": "meridian.def.net_revenue", "definition_version": 1, "unanswerable": false, "note": "13 invoices in period; 1 return (R002); gross minus discount."}, "raw": "```json\n{\n \"answer\": 109343,\n \"definition_id\": \"meridian.def.net_revenue\",\n \"definition_version\": 1,\n \"unanswerable\": false,\n \"note\": \"13 invoices in period; 1 return (R002); gross minus discount.\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-05", "cond": "full_catalogue", "rep": 0, "latency_s": 47.47, "prompt_chars": 21420, "context_chars": 10332, "answer_correct": 1, "obj": {"answer": 104670, "definition_id": "meridian.def.net_revenue", "definition_version": 2, "unanswerable": false, "note": "Period 2026-05-01 to 2026-06-30, v2 applies from 2026-05-01"}, "raw": "```json\n{\"answer\": 104670, \"definition_id\": \"meridian.def.net_revenue\", \"definition_version\": 2, \"unanswerable\": false, \"note\": \"Period 2026-05-01 to 2026-06-30, v2 applies from 2026-05-01\"}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "TEMP-01", "cond": "keyword_top5", "rep": 0, "latency_s": 57.38, "prompt_chars": 13249, "context_chars": 2178, "answer_correct": 1, "obj": {"answer": 21, "definition_id": "meridian.def.active_customer", "definition_version": 1, "unanswerable": false, "note": "Version 1 applies; interval (2025-02-28, 2026-02-28] with window 365 days."}, "raw": "```json\n{\n \"answer\": 21,\n \"definition_id\": \"meridian.def.active_customer\",\n \"definition_version\": 1,\n \"unanswerable\": false,\n \"note\": \"Version 1 applies; interval (2025-02-28, 2026-02-28] with window 365 days.\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-01", "cond": "full_catalogue", "rep": 0, "latency_s": 81.82, "prompt_chars": 21403, "context_chars": 10332, "answer_correct": 1, "obj": {"answer": 18, "definition_id": "meridian.def.active_customer", "definition_version": 2, "unanswerable": false, "note": "18 customers with non-cancelled orders in (2026-03-04, 2026-08-31]"}, "raw": "```json\n{\n \"answer\": 18,\n \"definition_id\": \"meridian.def.active_customer\",\n \"definition_version\": 2,\n \"unanswerable\": false,\n \"note\": \"18 customers with non-cancelled orders in (2026-03-04, 2026-08-31]\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-05", "cond": "full_catalogue", "rep": 1, "latency_s": 33.61, "prompt_chars": 21420, "context_chars": 10332, "answer_correct": 1, "obj": {"answer": 104670, "definition_id": "meridian.def.net_revenue", "definition_version": 2, "unanswerable": false, "note": "Invoices 2026-05-01 to 2026-06-30, minus returns in period."}, "raw": "```json\n{\n \"answer\": 104670,\n \"definition_id\": \"meridian.def.net_revenue\",\n \"definition_version\": 2,\n \"unanswerable\": false,\n \"note\": \"Invoices 2026-05-01 to 2026-06-30, minus returns in period.\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "TEMP-01", "cond": "full_catalogue", "rep": 0, "latency_s": 37.93, "prompt_chars": 21403, "context_chars": 10332, "answer_correct": 1, "obj": {"answer": 21, "definition_id": "meridian.def.active_customer", "definition_version": 1, "unanswerable": false, "note": "Customers with non-cancelled orders dated 2025-03-01 to 2026-02-28."}, "raw": "```json\n{\n \"answer\": 21,\n \"definition_id\": \"meridian.def.active_customer\",\n \"definition_version\": 1,\n \"unanswerable\": false,\n \"note\": \"Customers with non-cancelled orders dated 2025-03-01 to 2026-02-28.\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-10", "cond": "keyword_top5", "rep": 1, "latency_s": 38.66, "prompt_chars": 13100, "context_chars": 2024, "answer_correct": 0, "obj": {"answer": 22, "definition_id": "meridian.def.on_time", "definition_version": 2, "unanswerable": false, "note": "Applied v2 effective 2026-02-01; 22 shipped with customer_confirm_date≤2026-08-31 met on-time criteria"}, "raw": "```json\n{\"answer\": 22, \"definition_id\": \"meridian.def.on_time\", \"definition_version\": 2, \"unanswerable\": false, \"note\": \"Applied v2 effective 2026-02-01; 22 shipped with customer_confirm_date≤2026-08-31 met on-time criteria\"}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-05", "cond": "keyword_top5", "rep": 0, "latency_s": 30.86, "prompt_chars": 13158, "context_chars": 2070, "answer_correct": 0, "obj": {"answer": 91590, "definition_id": "meridian.def.net_revenue", "definition_version": 2, "unanswerable": false, "note": "v2 applies from 2026-05-01. Invoices in period: 11. Returns: 1 (R006)."}, "raw": "```json\n{\n \"answer\": 91590,\n \"definition_id\": \"meridian.def.net_revenue\",\n \"definition_version\": 2,\n \"unanswerable\": false,\n \"note\": \"v2 applies from 2026-05-01. Invoices in period: 11. Returns: 1 (R006).\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "XORG-04", "cond": "keyword_top5", "rep": 0, "latency_s": 61.17, "prompt_chars": 13116, "context_chars": 1992, "answer_correct": 1, "obj": {"answer": 11, "definition_id": "fed.C-118.on_time", "definition_version": null, "unanswerable": false, "note": "14 C-118 shipments delivered by 2026-08-31; 11 had carrier_scan_date on or before promised (or revised) delivery date."}, "raw": "```json\n{\n \"answer\": 11,\n \"definition_id\": \"fed.C-118.on_time\",\n \"definition_version\": null,\n \"unanswerable\": false,\n \"note\": \"14 C-118 shipments delivered by 2026-08-31; 11 had carrier_scan_date on or before promised (or revised) delivery date.\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-07", "cond": "self_selected", "rep": 1, "latency_s": 24.88, "prompt_chars": 11419, "context_chars": 351, "answer_correct": 1, "obj": {"answer": 50, "definition_id": "meridian.def.delivered", "definition_version": 1, "unanswerable": false, "note": "Shipments with customer_confirm_date on/before 2026-08-31 counted."}, "raw": "```json\n{\n \"answer\": 50,\n \"definition_id\": \"meridian.def.delivered\",\n \"definition_version\": 1,\n \"unanswerable\": false,\n \"note\": \"Shipments with customer_confirm_date on/before 2026-08-31 counted.\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-12", "cond": "keyword_top5", "rep": 1, "latency_s": 27.74, "prompt_chars": 13090, "context_chars": 1968, "answer_correct": 1, "obj": {"answer": 139070, "definition_id": "meridian.def.order_value", "definition_version": 1, "unanswerable": false, "note": "15 non-cancelled orders 2026-05-01 to 2026-07-31; value = gross minus discount"}, "raw": "```json\n{\n \"answer\": 139070,\n \"definition_id\": \"meridian.def.order_value\",\n \"definition_version\": 1,\n \"unanswerable\": false,\n \"note\": \"15 non-cancelled orders 2026-05-01 to 2026-07-31; value = gross minus discount\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "TEMP-07", "cond": "keyword_top5", "rep": 1, "latency_s": 40.92, "prompt_chars": 13158, "context_chars": 2070, "answer_correct": 1, "obj": {"answer": 109343, "definition_id": "meridian.def.net_revenue", "definition_version": 1, "unanswerable": false, "note": "Period 2026-01-01 to 2026-03-31 uses v1; Q1 invoices minus R002 return."}, "raw": "```json\n{\"answer\": 109343, \"definition_id\": \"meridian.def.net_revenue\", \"definition_version\": 1, \"unanswerable\": false, \"note\": \"Period 2026-01-01 to 2026-03-31 uses v1; Q1 invoices minus R002 return.\"}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "TEMP-01", "cond": "self_selected", "rep": 1, "latency_s": 48.05, "prompt_chars": 11562, "context_chars": 491, "answer_correct": 0, "obj": {"answer": 20, "definition_id": "meridian.def.active_customer", "definition_version": 1, "unanswerable": false, "note": "Applied window_days=365; interval (2025-02-28, 2026-02-28]"}, "raw": "```json\n{\n \"answer\": 20,\n \"definition_id\": \"meridian.def.active_customer\",\n \"definition_version\": 1,\n \"unanswerable\": false,\n \"note\": \"Applied window_days=365; interval (2025-02-28, 2026-02-28]\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "XORG-09", "cond": "self_selected", "rep": 1, "latency_s": 33.81, "prompt_chars": 11706, "context_chars": 457, "answer_correct": 1, "obj": {"answer": 9, "definition_id": "halden.def.business_day", "definition_version": 1, "unanswerable": false, "note": "Order O006 2025-12-30 to shipment scan 2026-01-14; 9 weekdays excluding 2026-01-01 and 2026-01-06 holidays and weekends."}, "raw": "```json\n{\n \"answer\": 9,\n \"definition_id\": \"halden.def.business_day\",\n \"definition_version\": 1,\n \"unanswerable\": false,\n \"note\": \"Order O006 2025-12-30 to shipment scan 2026-01-14; 9 weekdays excluding 2026-01-01 and 2026-01-06 holidays and weekends.\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-01", "cond": "self_selected", "rep": 1, "latency_s": 71.11, "prompt_chars": 11559, "context_chars": 488, "answer_correct": 1, "obj": {"answer": 18, "definition_id": "meridian.def.active_customer", "definition_version": 2, "unanswerable": false, "note": "Window 2026-03-05 to 2026-08-31; cancelled orders excluded."}, "raw": "```json\n{\n \"answer\": 18,\n \"definition_id\": \"meridian.def.active_customer\",\n \"definition_version\": 2,\n \"unanswerable\": false,\n \"note\": \"Window 2026-03-05 to 2026-08-31; cancelled orders excluded.\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-01", "cond": "full_catalogue", "rep": 1, "latency_s": 58.02, "prompt_chars": 21403, "context_chars": 10332, "answer_correct": 1, "obj": {"answer": 18, "definition_id": "meridian.def.active_customer", "definition_version": 2, "unanswerable": false, "note": "Counted customers with non-cancelled orders dated 2026-03-05 to 2026-08-31."}, "raw": "```json\n{\n \"answer\": 18,\n \"definition_id\": \"meridian.def.active_customer\",\n \"definition_version\": 2,\n \"unanswerable\": false,\n \"note\": \"Counted customers with non-cancelled orders dated 2026-03-05 to 2026-08-31.\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "XORG-04", "cond": "self_selected", "rep": 1, "latency_s": 60.21, "prompt_chars": 11800, "context_chars": 676, "answer_correct": 0, "obj": {"answer": 9, "definition_id": "meridian.def.on_time", "definition_version": 2, "unanswerable": false, "note": "9 of 15 delivered C-118 Halden shipments met promised/revised delivery date"}, "raw": "```json\n{\n \"answer\": 9,\n \"definition_id\": \"meridian.def.on_time\",\n \"definition_version\": 2,\n \"unanswerable\": false,\n \"note\": \"9 of 15 delivered C-118 Halden shipments met promised/revised delivery date\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-07", "cond": "keyword_top5", "rep": 0, "latency_s": 28.18, "prompt_chars": 13003, "context_chars": 1935, "answer_correct": 1, "obj": {"answer": 50, "definition_id": "meridian.def.delivered", "definition_version": 1, "unanswerable": false, "note": "Shipments with customer_confirm_date on/before 2026-08-31"}, "raw": "```json\n{\n \"answer\": 50,\n \"definition_id\": \"meridian.def.delivered\",\n \"definition_version\": 1,\n \"unanswerable\": false,\n \"note\": \"Shipments with customer_confirm_date on/before 2026-08-31\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "XORG-01", "cond": "keyword_top5", "rep": 1, "latency_s": 30.05, "prompt_chars": 13108, "context_chars": 1992, "answer_correct": 0, "obj": {"answer": 17, "definition_id": "fed.C-118.delivered", "definition_version": 2, "unanswerable": false, "note": "Halden C-118 shipments with carrier_scan_date on/before 2026-08-31"}, "raw": "```json\n{\n \"answer\": 17,\n \"definition_id\": \"fed.C-118.delivered\",\n \"definition_version\": 2,\n \"unanswerable\": false,\n \"note\": \"Halden C-118 shipments with carrier_scan_date on/before 2026-08-31\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "TEMP-07", "cond": "self_selected", "rep": 1, "latency_s": 44.28, "prompt_chars": 11593, "context_chars": 505, "answer_correct": 1, "obj": {"answer": 109343, "definition_id": "meridian.def.net_revenue", "definition_version": 1, "unanswerable": false, "note": "Sum of (gross - discount) per invoices in period, less returns."}, "raw": "```json\n{\"answer\": 109343, \"definition_id\": \"meridian.def.net_revenue\", \"definition_version\": 1, \"unanswerable\": false, \"note\": \"Sum of (gross - discount) per invoices in period, less returns.\"}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "COMP-10", "cond": "keyword_top5", "rep": 0, "latency_s": 67.96, "prompt_chars": 13100, "context_chars": 2024, "answer_correct": 1, "obj": {"answer": 23, "definition_id": "meridian.def.on_time", "definition_version": 2, "unanswerable": false, "note": "Applied v2 (2026-02-01+): on/before revised_promised_date or promised_date. 50 shipments had customer confirmation by 2026-08-31; 23 were on-time."}, "raw": "```json\n{\n \"answer\": 23,\n \"definition_id\": \"meridian.def.on_time\",\n \"definition_version\": 2,\n \"unanswerable\": false,\n \"note\": \"Applied v2 (2026-02-01+): on/before revised_promised_date or promised_date. 50 shipments had customer confirmation by 2026-08-31; 23 were on-time.\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "XORG-01", "cond": "self_selected", "rep": 0, "latency_s": 94.02, "prompt_chars": 11469, "context_chars": 353, "answer_correct": 0, "obj": {"answer": 16, "definition_id": "halden.def.delivered", "definition_version": 2, "unanswerable": false, "note": "C-118 shipments with carrier_scan_date on/before 2026-08-31"}, "raw": "```json\n{\"answer\": 16, \"definition_id\": \"halden.def.delivered\", \"definition_version\": 2, \"unanswerable\": false, \"note\": \"C-118 shipments with carrier_scan_date on/before 2026-08-31\"}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "XORG-04", "cond": "self_selected", "rep": 0, "latency_s": 100.2, "prompt_chars": 11800, "context_chars": 676, "answer_correct": 1, "obj": {"answer": 11, "definition_id": "meridian.def.on_time", "definition_version": 2, "unanswerable": false, "note": "18 C-118 HAL shipments delivered by 2026-08-31; 11 on-time per revised or promised dates."}, "raw": "```json\n{\n \"answer\": 11,\n \"definition_id\": \"meridian.def.on_time\",\n \"definition_version\": 2,\n \"unanswerable\": false,\n \"note\": \"18 C-118 HAL shipments delivered by 2026-08-31; 11 on-time per revised or promised dates.\"\n}\n```\n"} {"track": "B2", "stage": "answer", "model": "claude-haiku-4-5-20251001", "task": "XORG-04", "cond": "keyword_top5", "rep": 1, "latency_s": 135.37, "prompt_chars": 13116, "context_chars": 1992, "answer_correct": 0, "obj": {"answer": 10, "definition_id": "fed.C-118.on_time", "definition_version": null, "unanswerable": false, "note": "10 of 16 C-118 orders delivered on-time by 2026-08-31"}, "raw": "```json\n{\"answer\": 10, \"definition_id\": \"fed.C-118.on_time\", \"definition_version\": null, \"unanswerable\": false, \"note\": \"10 of 16 C-118 orders delivered on-time by 2026-08-31\"}\n```\n"}