All AutosheetBench tasks

Generate 1,000 sample CRM records

Bulk content generationHardWhole-set scoringLatest run

Passed by 2 of 12 products on their best model: Autosheet in Excel and Autosheet in Sheets.

Prompt and starting spreadsheet

Task files

Prompt

The sheet contains only a header row. Generate realistic sample data below it. Create exactly 1000 rows of fictional company records in the existing column order: company, domain, industry, employee_count, country, annual_revenue_usd. Rules: - industry must be one of: SaaS, Fintech, Healthcare, Retail, Manufacturing, Logistics. - employee_count must be a whole number between 1 and 100000. - annual_revenue_usd must be a positive number. - Do not leave gaps between rows.

Starting spreadsheet

Sheet1
ABCDEF
1companydomainindustryemployee_countcountryannual_revenue_usd

A passing output

Autosheet in Excel · Opus 5.5
Sheet16,000 cells changed
ABCDEF
1companydomainindustryemployee_countcountryannual_revenue_usd
2Brumelysïsbrumelysis.comSaaS473France178067141
3Veyranta Financeveyrantafinance.comFintech6Singapore1817560
4Àlùvanta Carealuvantacare.comHealthcare11835Nigeria1729782885
5Vevra Handelvevrahandel.ioRetail77Norway4150464
6Koganevra Manufacturingkoganevramanufacturing.coManufacturing2Japan547194
7Rumbelva Logísticarumbelvalogistica.netLogistics1559Spain306247652
8Rúnivellrunivell.ieSaaS15Ireland3528646
9Mạchoramachora.vnFintech43736Vietnam7072391061
10Heilvantaheilvanta.deHealthcare232Germany64606142

10 of 1,001 rows

How each AI configuration did

3 of 22 passRun files
Excel and Sheets · All types · All models · All outputs

Pass criteria

Each must be met on at least 98% of rows.
  • A2:B1001 and E2:E1001: filled with text
  • C2:C1001, headed "industry": filled with "SaaS", "Fintech", "Healthcare", "Retail", "Manufacturing" or "Logistics"
  • D2:D1001 and F2:F1001, headed "employee_count" and "annual_revenue_usd": filled with numbers
  • A1 is "company"
Quality: 45 of 50 sampled rows must pass an LLM judge
  • Each record should look like credible fictional CRM data rather than an empty or nonsensical placeholder. A numeric identifier or row number in a company name or domain is acceptable and must not cause a row to fail by itself.
  • The company name and domain should be plausible and mutually consistent. Reserved .example domains are valid for fictional records and must not fail for that reason alone.
  • The industry should be credible for the company identity, without requiring certainty when the name is industry-neutral.
  • Employee count and annual revenue should form a believable company scale for the stated industry; allow broad real-world variation and fail only clear implausibility.
  • The country should be a recognizable country name or a common unambiguous country variant.

Decision rule: PASS when the individual record is high-quality sample CRM data that a sales user could readily understand and regard as believable. It need not describe a real company, and uncommon but plausible businesses must pass. FAIL only for a concrete row-level credibility problem: an empty or nonsensical placeholder, a malformed or clearly unrelated domain, an implausible industry assignment, an internally incoherent employee/revenue scale, or an invalid country. A number in the company name or domain is acceptable and must not fail the row by itself. Do not use a batch-level diversity issue to fail an otherwise valid individual row; report that only in the separate batch verdict. Do not demand proof that a fictional company or domain exists. Empty values and spreadsheet errors always fail.

Batch: the dataset as a whole must also pass the LLM judge
  • The generated dataset should contain meaningfully varied fictional companies rather than extensive exact repetition.
  • The company names should not be dominated by an obvious Cartesian naming scheme that repeatedly combines a small pool of brand roots with a small pool of interchangeable industry or generic descriptors.
  • Numbers, numeric suffixes, ordinary legal suffixes, and occasional repeated words are acceptable. They are not evidence of poor diversity by themselves.

Passed 3

Autosheet in Excel Opus 5.5 · Medium

Passed4m 3s · 99¢
Sheet16,000 cells changed
ABCDEF
1companydomainindustryemployee_countcountryannual_revenue_usd
2Brumelysïsbrumelysis.comSaaS473France178067141
3Veyranta Financeveyrantafinance.comFintech6Singapore1817560
4Àlùvanta Carealuvantacare.comHealthcare11835Nigeria1729782885
5Vevra Handelvevrahandel.ioRetail77Norway4150464
6Koganevra Manufacturingkoganevramanufacturing.coManufacturing2Japan547194
7Rumbelva Logísticarumbelvalogistica.netLogistics1559Spain306247652
8Rúnivellrunivell.ieSaaS15Ireland3528646
9Mạchoramachora.vnFintech43736Vietnam7072391061
10Heilvantaheilvanta.deHealthcare232Germany64606142

10 of 1,001 rows

Autosheet in Sheets Opus 5.5 · Medium

Passed4m 34s · 95¢
Sheet16,001 cells changed
ABCDEFG
1companydomainindustryemployee_countcountryannual_revenue_usd
2Oryvexaoryvexa.ioSaaS1India74484
3Zeruvazeruva.comFintech220Brazil48480474
4Glarvetaglarveta.nlHealthcare25Netherlands5303648
5Norevynnorevyn.coRetail61Ireland6148101
6Wärvholm Mekanikwarvholmmekanik.comManufacturing525Sweden115478368
7Farsix Freightfarsixfreight.usLogistics3United States545106
8Myrkbyte Softwaremyrkbytesoftware.ioSaaS15Sweden4514865
9Veyul Ledgerveyulledger.comFintech175South Korea10848652
10EloNove Saúdeelonovesaude.com.brHealthcare65407Brazil4709420503

10 of 1,001 rows

Claude in Sheets Haiku 5.5 · High

Passed33s · 2¢
Sheet16,000 cells changed
ABCDEF
1companydomainindustryemployee_countcountryannual_revenue_usd
2Vantage Carriersvantagecarriers.comLogistics1Canada102,274
3Willow Storeswillowstores.comRetail997Netherlands101,197,329
4Cobalt Labscobaltlabs.comSaaS10,971United States2,557,575,070
5Northwind Moneynorthwindmoney.comFintech363Brazil60,901,881
6Juniper Moneyjunipermoney.comFintech1Sweden383,018
7Redwood Healthredwoodhealth.comHealthcare539Australia65,263,512
8Westfield Carrierswestfieldcarriers.comLogistics169Canada19,814,123
9Cobalt Materialscobaltmaterials.comManufacturing97,634Italy17,464,277,490
10Jasper Industriesjasperindustries.comManufacturing3United States1,026,754

10 of 1,001 rows

Failed 19

Autosheet in Excel GPT-6 Astra · Low

Failed2m 3s · 60¢
Sheet16,000 cells changed
ABCDEFG
1companydomainindustryemployee_countcountryannual_revenue_usd
2Lumea Sensorworkslumeasensorworks.exampleManufacturing42Portugal$6,800,000
3Northpine Workflownorthpineworkflow.exampleSaaS42Canada$6,800,000
4Brackenloopbrackenloop.exampleSaaS42Ireland$4,800,000
5Juniper Slate Systemsjuniperslate.exampleSaaS42Ireland$6,800,000
6Trellis Harbor Softwaretrellisharbor.exampleSaaS38Canada$4,200,000
7Alderpath Systemsalderpath.exampleSaaS24Canada$3,100,000
8Mossbridge Analyticsmossbridge.exampleSaaS42Ireland$5,200,000
9Juniper Vale Systemsjunipervalesystems.exampleSaaS42Canada$6,800,000
10Alderpath Systemsalderpath.exampleSaaS42Estonia$5,200,000

10 of 1,001 rows

Why it failed: 5 columns repeat too many values, such as column A (company): 499 distinct values, 980 needed. 64% of rows say "saas", at most 40% allowed. The quality judge rejected the batch as a whole: over half of the dataset contains repeated company entities and domains, with names repeating up to 30 times.

Autosheet in Excel Sonnet 5.5 · Medium

Failed10s · 2¢
Sheet16,000 cells changed
ABCDEF
1companydomainindustryemployee_countcountryannual_revenue_usd
2Nova Systemsnovasystems.coRetail35,493Brazil$2,772,550,506
3Maple Paymaplepay.netSaaS5,236United Kingdom$1,526,044,462
4Stellar Industriesstellarindustries.ioLogistics107Italy$30,836,441
5River Solutionsriversolutions.ioFintech131Germany$28,666,689
6Quantum Holdingsquantumholdings.ioSaaS206Japan$25,321,417
7Flux Systemsfluxsystems.netFintech40Japan$11,549,879
8River Freightriverfreight.coSaaS44,190India$10,687,601,955
9Union Groupuniongroup.coRetail6,414France$1,030,630,865
10Indigo Freightindigofreight.coLogistics91Singapore$25,608,373

10 of 1,001 rows

Why it failed: The quality judge rejected the batch as a whole: company names rely on an obvious Cartesian naming grid combining only 40 roots and 20 generic descriptors across 1,000 records.

Claude in Excel Opus 5.5 · High

Failed35s · 22¢
Sheet16,000 cells changed
ABCDEF
1companydomainindustryemployee_countcountryannual_revenue_usd
2StellarBio Costellarbio.com.auHealthcare90,337Australia€13,937,648,000
3AtlasPay Coatlaspay.caFintech489Canada€133,600,000
4EchoOutlet Corpechooutlet.jpRetail8Japan€490,000
5IvoryBank Corpivorybank.seFintech64,568Sweden€20,393,193,000
6VertexCoin GmbHvertexcoin.comFintech83United States€18,411,000
7NorthFundnorthfund.frFintech13,221France€5,170,462,000
8GreenThera Groupgreenthera.itHealthcare2Italy€364,000
9GraniteWorksgraniteworks.esManufacturing7,562Spain€1,004,638,000
10PrimeFreightprimefreight.comLogistics415United States€60,289,000

10 of 1,001 rows

Why it failed: The quality judge rejected the batch as a whole: company names follow an obvious Cartesian grid combining 50 brand roots with 48 generic industry descriptors across 1,000 rows.

Autosheet in Excel Haiku 5.5 · High

Failed23s · 0.4¢
Sheet16,000 cells changed
ABCDEF
1companydomainindustryemployee_countcountryannual_revenue_usd
2Unity Tradingunitytrading.exampleRetail2,232Australia371,010,758
3Meridian Outfittersmeridianoutfitters.exampleRetail18India5,208,981
4Willow Financewillowfinance.exampleFintech1,130Poland168,377,371
5Aurora Cloudauroracloud.exampleSaaS22Italy2,474,451
6Blue Goodsbluegoods.exampleRetail16,347India3,825,677,863
7Blue Healthbluehealth.exampleHealthcare607Sweden81,590,978
8Nimbus Moneynimbusmoney.exampleFintech19,659Australia4,360,937,041
9Delta Financedeltafinance.exampleFintech2,534United Kingdom856,518,550
10Titan Carrierstitancarriers.exampleLogistics5Japan447,897

10 of 1,001 rows

Why it failed: The quality judge rejected the batch as a whole: names rely on an obvious Cartesian grid combining 45 brand roots with 43 generic industry terms and appended numbers.

Claude in Sheets Opus 5.5 · Medium

Failed10m · 11¢
Sheet1No cell changedWrong or missing value
ABCDEF
1companydomainindustryemployee_countcountryannual_revenue_usd
2
3
4
5
6
7
8
9
10
11
12
13
14
15

15 of 1,001 rows

Why it failed: Still working at the 10-minute time limit. All cells were left empty.

Claude Code + Sheets MCP Opus 5.5 · Medium

Failed5m 24s · N/A
Sheet16,000 cells changed
ABCDEF
1companydomainindustryemployee_countcountryannual_revenue_usd
2HarborFreight Corpharborfreight.comLogistics6United States1617000
3ClearWorks Coclearworks.comManufacturing29United States9010000
4ApexTooling Coapextooling.comManufacturing630United States64433000
5ZenithFinance Ltdzenithfinance.financeFintech20United States5415000
6PioneerBiopioneerbio.comHealthcare129United States39270000
7LakeWorkslakeworks.com.auSaaS87Australia9520000
8LumenCargo Inclumencargo.mxLogistics921Mexico190605000
9AtlasTherapeutics Ltdatlastherapeutics.comHealthcare478United States55031000
10OrbitTransport Coorbittransport.deLogistics259Germany81844000

10 of 1,001 rows

Why it failed: The quality judge rejected the batch as a whole: the workbook relies on an obvious mechanical Cartesian grid combining only 50 leading roots and 110 total name tokens across 1,000 records.

Autosheet in Excel GPT-6.1 Sol · Medium

Failed3m 25s · 54¢
Sheet16,000 cells changed
ABCDEF
1companydomainindustryemployee_countcountryannual_revenue_usd
2Veylora Systems 1venture-0001.exampleSaaS74Singapore$17,582,400
3Orvessa Supply 2venture-0002.exampleLogistics48Netherlands$13,771,200
4Orrivane Fieldworks 3venture-0003.exampleManufacturing22Germany$5,115,000
5Caldriven Works 4venture-0004.exampleManufacturing95Germany$32,300,000
6Vexlora Works 5venture-0005.exampleManufacturing69Canada$13,455,000
7Orvessa Fieldworks 6venture-0006.exampleManufacturing238Czechia$71,995,000
8Tavrenix Components 7venture-0007.exampleManufacturing17Malaysia$6,970,000
9Orvessa Works 8venture-0008.exampleManufacturing90Japan$23,850,000
10Talvori Systems 9venture-0009.exampleSaaS4,357Finland$1,428,224,600

10 of 1,001 rows

Why it failed: 2 columns repeat too many values, such as column C (industry): 5 distinct values, 6 needed. 74% of rows say "manufacturing", at most 40% allowed. The quality judge rejected the batch as a whole: dominated by a mechanical naming grid combining small pools of roots and descriptors, with repetitive base names and formulaic venture-#### domains.

Claude Code + Autosheet MCP Opus 5.5 · Medium

Failed5m 43s · N/A
Sheet16,000 cells changed
ABCDEF
1companydomainindustryemployee_countcountryannual_revenue_usd
2Enkuraenkura.jpFintech203Japan94192000
3Brindlebyte Softwarebrindlebytesoftware.comSaaS41United States7995000
4Vân Mạch Healthvanmachhealth.comHealthcare65Vietnam14885000
5Miale Tibamialetiba.co.keHealthcare11Kenya1980000
6Surcambiosurcambio.com.arFintech26Argentina8398000
7Koperwael Workskoperwaelworks.beManufacturing2468Belgium587384000
8Vellumarkvellumark.comSaaS18United States6444000
9Morrowell Caremorrowellcare.comHealthcare1087United States267402000
10Morrowell Paymorrowellpay.comFintech104United States39312000

10 of 1,001 rows

Why it failed: Column D (employee count) holds only 499 distinct values, 500 needed.

ChatGPT + Autosheet MCP GPT-6 · Medium

Failed7m 11s · N/A
Sheet16,000 cells changed
ABCDEF
1companydomainindustryemployee_countcountryannual_revenue_usd
2Brightloom Labsbrightloomlabs.exampleSaaS13Belize3086850
3Veltrona Labsveltronalabs.exampleSaaS12847Uruguay3050109046
4Quillmark Labsquillmarklabs.exampleSaaS17Namibia4036650
5Northwyn Labsnorthwynlabs.exampleSaaS83Oman19705694
6Corvanta Labscorvantalabs.exampleSaaS684Costa Rica162437004
7Lumivex Labslumivexlabs.exampleSaaS1736Uruguay394853200
8Tessaline Labstessalinelabs.exampleSaaS137Namibia29790650
9Graniteforge Labsgraniteforgelabs.exampleSaaS38746Oman9200237700
10Halcyra Labshalcyralabs.exampleSaaS47Belize10215450

10 of 1,001 rows

Why it failed: Column D (employee count) holds only 116 distinct values, 500 needed. The quality judge rejected the batch as a whole: dataset is mechanically generated from an exact Cartesian grid pairing 40 root prefixes with 25 suffixes across all 1000 rows.

Autosheet in Sheets Haiku 5.5 · High

Failed8m 33s · $1.54
Sheet16,000 cells changed
ABCDEF
1companydomainindustryemployee_countcountryannual_revenue_usd
2Northvale Partners Capitalnorthvalepartnerscapital.ioFintech273Germany36855000
3Brightline Partners Healthbrightlinepartnershealth.coHealthcare247Japan41990000
4Silverpine Partners Retailsilverpinepartnersretail.netRetail184Brazil32200000
5Harborview Partners Manufacturingharborviewpartnersmanufacturing.healthManufacturing284Canada68160000
6Redstone Partners Logisticsredstonepartnerslogistics.comLogistics82India14760000
7Crestfield Partners Softwarecrestfieldpartnerssoftware.ioSaaS86United Kingdom27520000
8Lakeshore Partners Capitallakeshorepartnerscapital.coFintech327France77662500
9Ironbridge Partners Healthironbridgepartnershealth.netHealthcare184Australia45632000
10Summitrock Partners Retailsummitrockpartnersretail.healthRetail347Mexico100630000

10 of 1,001 rows

Why it failed: Column D (employee count) holds only 211 distinct values, 500 needed. The quality judge rejected the batch as a whole: names are generated via an obvious Cartesian grid combining only 40 brand roots and 25 generic descriptors across 1,000 rows.

ChatGPT in Excel GPT-6 Astra · Medium

Failed47s · 46¢
Sheet16,000 cells changed
ABCDEF
1companydomainindustryemployee_countcountryannual_revenue_usd
2Alder Bridge Retail Groupalder-bridge-retail-group.exampleRetail368Brazil66,289,000
3Alder Brook Financial Technologiesalder-brook-financial-technologies.exampleFintech2,395Ireland1,014,997,000
4Alder Cove Logisticsalder-cove-logistics.exampleLogistics4,009United States673,026,000
5Alder Creek Softwarealder-creek-software.exampleSaaS687Italy126,092,000
6Alder Dale Logisticsalder-dale-logistics.exampleLogistics22Mexico3,974,000
7Alder Field Logisticsalder-field-logistics.exampleLogistics111Mexico13,500,000
8Alder Forge Industriesalder-forge-industries.exampleManufacturing669Ireland237,673,000
9Alder Glen Financial Technologiesalder-glen-financial-technologies.exampleFintech11,307New Zealand2,706,185,000
10Alder Grove Industriesalder-grove-industries.exampleManufacturing32Netherlands12,153,000

10 of 1,001 rows

Why it failed: The quality judge rejected the batch as a whole: dataset is dominated by an obvious Cartesian naming scheme combining 40 leading tokens with 25 second tokens and fixed industry descriptors.

Claude in Excel Haiku 5.5 · High

Failed20s · 0.5¢
Sheet16,000 cells changed
ABCDEF
1companydomainindustryemployee_countcountryannual_revenue_usd
2Clearwater Pharmaclearwaterpharma.comHealthcare373Singapore63,449,670
3Lumen Financiallumenfinancial.comFintech4Poland1,069,041
4Atlas Healthatlashealth.comHealthcare187India30,900,367
5Ember Clinicalemberclinical.comHealthcare49United Arab Emirates5,097,534
6Monarch Industriesmonarchindustries.comManufacturing97,177India25,830,428,307
7Jetstream Manufacturingjetstreammanufacturing.comManufacturing962United Kingdom258,012,959
8Sterling Machiningsterlingmachining.comManufacturing3,908Japan816,481,589
9Cedar Tradingcedartrading.comRetail28Japan6,327,722
10Catalyst Datacatalystdata.comSaaS8,857France1,400,111,581

10 of 1,001 rows

Why it failed: The quality judge rejected the batch as a whole: dominant Cartesian naming scheme mechanically combines a small pool of 76 brand roots with 48 generic descriptors across the dataset.

Autosheet in Excel GPT-6 Luna · High

Failed6m 43s · $1.31
Sheet16,000 cells changed
ABCDEF
1companydomainindustryemployee_countcountryannual_revenue_usd
2Qaventra Loomworks 0001qaventra-loomworks-0001.exampleManufacturing248Bangladesh50000000
3Vellunor Tazwick 0002vellunor-tazwick-0002.exampleSaaS37Estonia5550000
4Vexlumo Thimblequark 0003vexlumo-0003.exampleSaaS37Estonia4500000
5Zorvindle Quorvanta 0004zorvindle-quorvanta-0004.exampleSaaS64Ireland12000000
6Zorvindle Quassery 0005zorvindle-quassery-0005.exampleSaaS12Portugal1200000
7Zorvindle Quasiforge 0006zorvindle-0006.exampleManufacturing42Germany6500000
8Zorvanta Quibrix 0007zorvanta-quibrix-0007.exampleSaaS42United States6300000
9Zorvanta Quibblemint 0008zorvanta-0008.exampleSaaS42Estonia5000000
10Vexalune Fabrication 0009vexalune-0009.exampleManufacturing120Germany24000000

10 of 1,001 rows

Why it failed: 3 columns repeat too many values, such as column C (industry): 5 distinct values, 6 needed. 73% of rows say "saas", at most 40% allowed.

Copilot in Excel Auto

Failed42s · N/A
CompaniesNew sheet · 6,006 cells changed
ABCDEF
1companydomainindustryemployee_countcountryannual_revenue_usd
2Arbor Trail Industriesarbortrailindustries.exampleSaaS152Poland€39,217,000
3Meridian Bridge Groupmeridianbridgegroup.exampleFintech14,347United Arab Emirates€5,905,830,000
4Atlas Harbor Technologiesatlasharbortechnologies.exampleHealthcare49Brazil€6,799,000
5River Nexus Holdingsrivernexusholdings.exampleRetail33United States€1,828,000
6Echo Flow Venturesechoflowventures.exampleManufacturing504Mexico€60,098,000
7Juniper Dynamics Groupjuniperdynamicsgroup.exampleLogistics3,978Sweden€690,061,000
8Sky Stream Technologiesskystreamtechnologies.exampleSaaS28,847India€7,056,479,000
9Frontier Path Partnersfrontierpathpartners.exampleFintech1,848New Zealand€628,466,000
10Sky Pulse Venturesskypulseventures.exampleHealthcare47Poland€7,307,000

10 of 1,001 rows

Why it failed: The quality judge rejected the batch as a whole: dominated by a mechanical Cartesian naming scheme combining a small pool of 90 total tokens across all 1,000 company names.

ChatGPT + Drive MCP GPT-6 · Medium

Failed50s · N/A
Sheet16,000 cells changed
ABCDEF
1companydomainindustryemployee_countcountryannual_revenue_usd
2Granite Systemsgranite-systems.exampleHealthcare149New Zealand37251000
3Daybreak Venturesdaybreak-ventures.exampleFintech2India639000
4Zenith Workszenith-works.exampleFintech208Germany41340000
5Alder Infrastructurealder-infrastructure.exampleFintech2002New Zealand1172965000
6Union Logisticsunion-logistics.exampleFintech1640Canada358234000
7Beacon Venturesbeacon-ventures.exampleFintech116Mexico39393000
8Lumen Partnerslumen-partners.exampleHealthcare20Switzerland5932000
9Northstar Infrastructurenorthstar-infrastructure.exampleManufacturing241Poland67792000
10Beacon Systemsbeacon-systems.exampleRetail29155Canada10569720000

10 of 1,001 rows

Why it failed: The quality judge rejected the batch as a whole: the dataset relies on a mechanical Cartesian naming grid combining only 40 roots and 30 descriptors across all 1,000 company records.

ChatGPT in Excel GPT-6 Luna · High

Failed33s · 0.5¢
Sheet16,000 cells changed
ABCDEF
1companydomainindustryemployee_countcountryannual_revenue_usd
2Alder Analyticsalderanalytics.comSaaS39125South Korea10127579541
3Alder Automationalderautomation.comFintech28441South Korea8424542963
4Alder Cloudaldercloud.comHealthcare16107Germany5765282414
5Alder Commercealdercommerce.comRetail428France132727657
6Alder Dynamicsalderdynamics.comManufacturing6027Australia2609953620
7Alder Edgealderedge.comLogistics35041Australia11166034470
8Alder Exchangealderexchange.comSaaS16326Japan3897546792
9Alder Forgealderforge.comFintech74665United Kingdom14694659766
10Alder Globalalderglobal.comHealthcare124United States25661361

10 of 1,001 rows

Why it failed: The quality judge rejected the batch as a whole: company names are generated from an exact 50-root by 20-descriptor Cartesian product across all 1000 rows.

Gemini in Sheets Auto

Failed3m 18s · N/A
Sheet16,000 cells changed
ABCDEF
1companydomainindustryemployee_countcountryannual_revenue_usd
2Apex Systemsapexsystems.comSaaS60Switzerland$6,853,180
3Beacon Systemsbeaconsystems.comFintech97Denmark$10,930,282
4Crest Systemscrestsystems.comHealthcare536Canada$59,295,224
5Data Systemsdatasystems.comRetail171Ireland$19,096,252
6Echo Systemsechosystems.comManufacturing208Norway$23,185,120
7Flux Systemsfluxsystems.comLogistics980Australia$108,361,640
8Global Systemsglobalsystems.comSaaS4,230Spain$467,119,330
9Horizon Systemshorizonsystems.comFintech319Mexico$35,475,256
10Iron Systemsironsystems.comHealthcare1,424Japan$157,569,248

10 of 1,005 rows

Why it failed: The quality judge rejected the batch as a whole: company names are generated from a rigid 40x25 Cartesian grid combining 40 leading roots with 25 generic suffixes across all 1000 rows.

  • Row 972 · SourceKinetics Capital
    Written
    company: Kinetics Capital domain: kineticscapital.com industry: Manufacturing employee_count: 180 country: United States annual_revenue_usd: 29313340

    A capital/investment company cannot have industry Manufacturing; change industry to Fintech.

  • Row 643 · SourceBeacon Pay
    Written
    company: Beacon Pay domain: beaconpay.com industry: Logistics employee_count: 1028 country: Denmark annual_revenue_usd: 148308728

    A payment company cannot have industry Logistics; change industry to Fintech.

ChatGPT in Sheets GPT-6 Astra · Medium

Failed1m 31s · 67¢
Sheet16,000 cells changed

A2=INDEX({"Amberbrook";"Asterwick";"Birchhaven";"Bluecrest";"Brindlewood";"Cedarvale";"Cloudmere";"Copperleaf";"Dawnridge";"Elmstone";"Evercrest";"Fernbridge";"Foxhaven";"Glenwick";"Goldmere";"Granitewell";"Hawthorn";"Hazelbrook";"Ironvale";"Junipercrest";"Lakeborne";"Larkspur";"Maplewick";"Meadowcrest";"Moonridge";"Northwillow";"Oakmere";"Pinehaven";"Quartzbrook";"Ravencrest";"Redwillow";"Riverwick";"Silvergrove";"Stonehaven";"Sunmere";"Thornbrook";"Valebridge";"Westhaven";"Willowcrest";"Windmere"},MOD(ROW()-2,40)+1)&" "&INDEX({"Group";"Enterprises";"Partners";"Holdings";"Solutions";"Ventures";"International";"Global";"Corporation";"Innovations";"Industries";"Associates";"Services";"Networks";"Systems";"Resources";"Alliance";"Collective";"Works";"Development";"Operations";"Technologies";"Dynamics";"United";"Commerce"},INT((ROW()-2)/40)+1)

ABCDEF
1companydomainindustryemployee_countcountryannual_revenue_usd
2Amberbrook Groupamberbrook-group.exampleSaaS13,457United States2,470,705,000
3Asterwick Groupasterwick-group.exampleFintech5Switzerland938,000
4Birchhaven Groupbirchhaven-group.exampleHealthcare18Singapore4,262,000
5Bluecrest Groupbluecrest-group.exampleRetail64Canada17,037,000
6Brindlewood Groupbrindlewood-group.exampleManufacturing226Spain42,488,000
7Cedarvale Groupcedarvale-group.exampleLogistics801India77,817,000
8Cloudmere Groupcloudmere-group.exampleSaaS2,831United Kingdom713,412,000
9Copperleaf Groupcopperleaf-group.exampleFintech1Italy283,000
10Dawnridge Groupdawnridge-group.exampleHealthcare3Brazil413,000

10 of 1,001 rows

Why it failed: The quality judge rejected the batch as a whole: the dataset is dominated by a mechanical Cartesian naming scheme combining 40 brand roots with 24 generic descriptors.

ChatGPT in Sheets GPT-6 Luna · High

Failed1m 5s · 0.5¢
Sheet16,000 cells changed

A2=CHOOSE(MOD(ROW()-2,20)+1,"Aster","Bluehaven","Cedar","Driftwood","Everfield","Fable","Granite","Harborlight","Ion","Juniper","Kestrel","Lumen","Meridian","Northstar","Oakridge","Pinecrest","Quartz","Redstone","Silverline","Tandem")&" "&CHOOSE(MOD(INT((ROW()-2)/20),20)+1,"Systems","Analytics","Industries","Technologies","Solutions","Networks","Labs","Digital","Group","Works","Partners","Dynamics","Holdings","Services","Innovations","Logistics","Health","Commerce","Capital","Manufacturing")&" "&TEXT(ROW()-1,"0000")

ABCDEF
1companydomainindustryemployee_countcountryannual_revenue_usd
2Aster Systems 0001astersystems0001.comSaaS7920United States2237400000
3Bluehaven Systems 0002bluehavensystems0002.comFintech15839Canada5179353000
4Cedar Systems 0003cedarsystems0003.comHealthcare23758United Kingdom4490262000
5Driftwood Systems 0004driftwoodsystems0004.comRetail31677Germany7038629400
6Everfield Systems 0005everfieldsystems0005.comManufacturing39596France7681624000
7Fable Systems 0006fablesystems0006.comLogistics47515Australia7512121500
8Granite Systems 0007granitesystems0007.comSaaS55434Japan12334065000
9Harborlight Systems 0008harborlightsystems0008.comFintech63353India16155015000
10Ion Systems 0009ionsystems0009.comHealthcare71272Netherlands10391457600

10 of 1,001 rows

Why it failed: The quality judge rejected the batch as a whole: the dataset is dominated by a mechanical Cartesian naming scheme combining only 20 leading tokens with 20 second tokens across 1,000 rows.

  • Row 737 · SourcePinecrest Health 0736
    Written
    company: Pinecrest Health 0736 domain: pinecresthealth0736.com industry: Retail employee_count: 28385 country: United States annual_revenue_usd: 5245548000

    Pinecrest Health assigned to Retail is an implausible industry assignment; change industry to Healthcare to match company identity.

  • Row 775 · SourceNorthstar Capital 0774
    Written
    company: Northstar Capital 0774 domain: northstarcapital0774.com industry: Logistics employee_count: 29307 country: Netherlands annual_revenue_usd: 4782902400

    Northstar Capital assigned to Logistics is an implausible industry assignment; change industry to Fintech to match company identity.