Skip to main content

LAS

LAS maps orders to dispatches and determineswhich primary contractor receives the transport data

LAS splits even consolidated (mixed) loads by primary contractor in three stages: LAS pre-determination, LAS model determination and LAS second-pass determination. Ambiguous dispatches are also confirmed automatically in the second pass, so nothing is left for manual review.

  • LAS pre-determination
  • LAS model determination
  • LAS second-pass determination
LAS engineLogistics determination in commercial useLAS engineLogistics determination in commercial useSame question catalog · state builder · policy stepsSame question catalog · state builder · policy stepsLAS in-house model4 billion parameters · open weightsPortable to customer servers · vehicle unitsLAS in-house model4 billion parameters · open weightsPortable to customer servers · vehicle unitsNew weightsNew weightsConnected vehiclesLTS DTG trip signal44 CAN profiles by vehicle typeLTS DTG trip signal44 CAN profiles by vehicle typeContractor orders · dispatchNames · business relationshipsContractor orders · dispatchNames · business relationships1122334455Link trips · LAS determinationCollect wrong & 2nd-pass casesAnswer confirmedRetrain · held-out test · deployLink trips · LAS determinationCollect wrong & 2nd-pass casesAnswer confirmedRetrain · held-out test · deployConfirmed by official registry IDsConfirmed by official registry IDs1,060 pairs so far1,060 pairs so farDetermination casesDetermination cases

LAS measured · 180 dispatches

The LAS engine resolved 180 dispatches in 5.0 seconds,35× and 95× faster than two frontier language models

5.0 sTo resolve all 180 dispatches
35×Speed vs. Claude Fable 5.1176.1 s → 5.0 s
95×Speed vs. GPT-6 Astra477.8 s → 5.0 s
0.65 KRWCost per runOver 2,000× lower than both models
Time and cost to finish the same full job — 180 dispatchesNormalize primary-contractor order, dispatch and DTG trip CSVs → match dispatches to orders → allocate consolidated loads → total per primary contractor180 dispatches · median of 3 runs per engine · public API rates (checked 2026-09-29) · exchange rate KRW 1,359.4/USD (2026-09-29)Time: LAS = server processing time · LLMs = call round-trip time (end to end)Engine elapsed timePlayback at 60× · same speed from start to finish01 min2 min3 min4 min5 min6 min7 min8 minLAS engineCorrect in all 3 runsCumulative costClaude Fable 5.1Correct in all 3 runsCumulative costGPT-6 AstraCorrect in all 3 runsCumulative cost0:00.0Processing · 0.0 sProcessing · 0.0 sProcessing · 0.0 s0.00 KRW0.0 KRW0.0 KRW0:05.0Processing · 5.0 sProcessing · 5.0 s50.0 KRWProcessing · 5.0 s14.9 KRW0:10.0Finished the same job 2 times meanwhileProcessing · 10.0 s100.0 KRWProcessing · 10.0 s29.7 KRW0:15.0Finished the same job 3 times meanwhileProcessing · 15.0 s149.9 KRWProcessing · 15.0 s44.6 KRW0:20.0Finished the same job 4 times meanwhileProcessing · 20.0 s199.9 KRWProcessing · 20.0 s59.4 KRW0:25.0Finished the same job 5 times meanwhileProcessing · 25.0 s249.9 KRWProcessing · 25.0 s74.3 KRW0:30.0Finished the same job 6 times meanwhileProcessing · 30.0 s299.9 KRWProcessing · 30.0 s89.1 KRW0:35.0Finished the same job 7 times meanwhileProcessing · 35.0 s349.9 KRWProcessing · 35.0 s104.0 KRW0:40.0Finished the same job 8 times meanwhileProcessing · 40.0 s399.8 KRWProcessing · 40.0 s118.8 KRW0:45.0Finished the same job 9 times meanwhileProcessing · 45.0 s449.8 KRWProcessing · 45.0 s133.7 KRW0:50.0Finished the same job 10 times meanwhileProcessing · 50.0 s499.8 KRWProcessing · 50.0 s148.6 KRW0:55.0Finished the same job 11 times meanwhileProcessing · 55.0 s549.8 KRWProcessing · 55.0 s163.4 KRW1:00.0Finished the same job 12 times meanwhileProcessing · 1 min 00 s599.8 KRWProcessing · 1 min 00 s178.3 KRW1:05.0Finished the same job 13 times meanwhileProcessing · 1 min 05 s649.7 KRWProcessing · 1 min 05 s193.1 KRW1:10.0Finished the same job 14 times meanwhileProcessing · 1 min 10 s699.7 KRWProcessing · 1 min 10 s208.0 KRW1:15.0Finished the same job 15 times meanwhileProcessing · 1 min 15 s749.7 KRWProcessing · 1 min 15 s222.8 KRW1:20.0Finished the same job 16 times meanwhileProcessing · 1 min 20 s799.7 KRWProcessing · 1 min 20 s237.7 KRW1:25.0Finished the same job 17 times meanwhileProcessing · 1 min 25 s849.7 KRWProcessing · 1 min 25 s252.5 KRW1:30.0Finished the same job 18 times meanwhileProcessing · 1 min 30 s899.6 KRWProcessing · 1 min 30 s267.4 KRW1:35.0Finished the same job 19 times meanwhileProcessing · 1 min 35 s949.6 KRWProcessing · 1 min 35 s282.3 KRW1:40.0Finished the same job 20 times meanwhileProcessing · 1 min 40 s999.6 KRWProcessing · 1 min 40 s297.1 KRW1:45.0Finished the same job 21 times meanwhileProcessing · 1 min 45 s1,049.6 KRWProcessing · 1 min 45 s312.0 KRW1:50.0Finished the same job 22 times meanwhileProcessing · 1 min 50 s1,099.6 KRWProcessing · 1 min 50 s326.8 KRW1:55.0Finished the same job 23 times meanwhileProcessing · 1 min 55 s1,149.5 KRWProcessing · 1 min 55 s341.7 KRW2:00.0Finished the same job 24 times meanwhileProcessing · 2 min 00 s1,199.5 KRWProcessing · 2 min 00 s356.5 KRW2:05.0Finished the same job 25 times meanwhileProcessing · 2 min 05 s1,249.5 KRWProcessing · 2 min 05 s371.4 KRW2:10.0Finished the same job 26 times meanwhileProcessing · 2 min 10 s1,299.5 KRWProcessing · 2 min 10 s386.2 KRW2:15.0Finished the same job 27 times meanwhileProcessing · 2 min 15 s1,349.5 KRWProcessing · 2 min 15 s401.1 KRW2:20.0Finished the same job 28 times meanwhileProcessing · 2 min 20 s1,399.4 KRWProcessing · 2 min 20 s416.0 KRW2:25.0Finished the same job 29 times meanwhileProcessing · 2 min 25 s1,449.4 KRWProcessing · 2 min 25 s430.8 KRW2:30.0Finished the same job 30 times meanwhileProcessing · 2 min 30 s1,499.4 KRWProcessing · 2 min 30 s445.7 KRW2:35.0Finished the same job 31 times meanwhileProcessing · 2 min 35 s1,549.4 KRWProcessing · 2 min 35 s460.5 KRW2:40.0Finished the same job 32 times meanwhileProcessing · 2 min 40 s1,599.4 KRWProcessing · 2 min 40 s475.4 KRW2:45.0Finished the same job 33 times meanwhileProcessing · 2 min 45 s1,649.3 KRWProcessing · 2 min 45 s490.2 KRW2:50.0Finished the same job 34 times meanwhileProcessing · 2 min 50 s1,699.3 KRWProcessing · 2 min 50 s505.1 KRW2:55.0Finished the same job 35 times meanwhileProcessing · 2 min 55 s1,749.3 KRWProcessing · 2 min 55 s519.9 KRW3:00.0Finished the same job 36 times meanwhileProcessing · 3 min 00 s534.8 KRW3:05.0Finished the same job 37 times meanwhileProcessing · 3 min 05 s549.7 KRW3:10.0Finished the same job 38 times meanwhileProcessing · 3 min 10 s564.5 KRW3:15.0Finished the same job 39 times meanwhileProcessing · 3 min 15 s579.4 KRW3:20.0Finished the same job 40 times meanwhileProcessing · 3 min 20 s594.2 KRW3:25.0Finished the same job 41 times meanwhileProcessing · 3 min 25 s609.1 KRW3:30.0Finished the same job 42 times meanwhileProcessing · 3 min 30 s623.9 KRW3:35.0Finished the same job 43 times meanwhileProcessing · 3 min 35 s638.8 KRW3:40.0Finished the same job 44 times meanwhileProcessing · 3 min 40 s653.6 KRW3:45.0Finished the same job 45 times meanwhileProcessing · 3 min 45 s668.5 KRW3:50.0Finished the same job 46 times meanwhileProcessing · 3 min 50 s683.4 KRW3:55.0Finished the same job 47 times meanwhileProcessing · 3 min 55 s698.2 KRW4:00.0Finished the same job 48 times meanwhileProcessing · 4 min 00 s713.1 KRW4:05.0Finished the same job 49 times meanwhileProcessing · 4 min 05 s727.9 KRW4:10.0Finished the same job 50 times meanwhileProcessing · 4 min 10 s742.8 KRW4:15.0Finished the same job 51 times meanwhileProcessing · 4 min 15 s757.6 KRW4:20.0Finished the same job 52 times meanwhileProcessing · 4 min 20 s772.5 KRW4:25.0Finished the same job 53 times meanwhileProcessing · 4 min 25 s787.3 KRW4:30.0Finished the same job 54 times meanwhileProcessing · 4 min 30 s802.2 KRW4:35.0Finished the same job 55 times meanwhileProcessing · 4 min 35 s817.1 KRW4:40.0Finished the same job 56 times meanwhileProcessing · 4 min 40 s831.9 KRW4:45.0Finished the same job 57 times meanwhileProcessing · 4 min 45 s846.8 KRW4:50.0Finished the same job 58 times meanwhileProcessing · 4 min 50 s861.6 KRW4:55.0Finished the same job 59 times meanwhileProcessing · 4 min 55 s876.5 KRW5:00.0Finished the same job 60 times meanwhileProcessing · 5 min 00 s891.3 KRW5:05.0Finished the same job 61 times meanwhileProcessing · 5 min 05 s906.2 KRW5:10.0Finished the same job 62 times meanwhileProcessing · 5 min 10 s921.0 KRW5:15.0Finished the same job 63 times meanwhileProcessing · 5 min 15 s935.9 KRW5:20.0Finished the same job 64 times meanwhileProcessing · 5 min 20 s950.8 KRW5:25.0Finished the same job 65 times meanwhileProcessing · 5 min 25 s965.6 KRW5:30.0Finished the same job 66 times meanwhileProcessing · 5 min 30 s980.5 KRW5:35.0Finished the same job 67 times meanwhileProcessing · 5 min 35 s995.3 KRW5:40.0Finished the same job 68 times meanwhileProcessing · 5 min 40 s1,010.2 KRW5:45.0Finished the same job 69 times meanwhileProcessing · 5 min 45 s1,025.0 KRW5:50.0Finished the same job 70 times meanwhileProcessing · 5 min 50 s1,039.9 KRW5:55.0Finished the same job 71 times meanwhileProcessing · 5 min 55 s1,054.7 KRW6:00.0Finished the same job 72 times meanwhileProcessing · 6 min 00 s1,069.6 KRW6:05.0Finished the same job 73 times meanwhileProcessing · 6 min 05 s1,084.5 KRW6:10.0Finished the same job 74 times meanwhileProcessing · 6 min 10 s1,099.3 KRW6:15.0Finished the same job 75 times meanwhileProcessing · 6 min 15 s1,114.2 KRW6:20.0Finished the same job 76 times meanwhileProcessing · 6 min 20 s1,129.0 KRW6:25.0Finished the same job 77 times meanwhileProcessing · 6 min 25 s1,143.9 KRW6:30.0Finished the same job 78 times meanwhileProcessing · 6 min 30 s1,158.7 KRW6:35.0Finished the same job 79 times meanwhileProcessing · 6 min 35 s1,173.6 KRW6:40.0Finished the same job 80 times meanwhileProcessing · 6 min 40 s1,188.4 KRW6:45.0Finished the same job 81 times meanwhileProcessing · 6 min 45 s1,203.3 KRW6:50.0Finished the same job 82 times meanwhileProcessing · 6 min 50 s1,218.2 KRW6:55.0Finished the same job 83 times meanwhileProcessing · 6 min 55 s1,233.0 KRW7:00.0Finished the same job 84 times meanwhileProcessing · 7 min 00 s1,247.9 KRW7:05.0Finished the same job 85 times meanwhileProcessing · 7 min 05 s1,262.7 KRW7:10.0Finished the same job 86 times meanwhileProcessing · 7 min 10 s1,277.6 KRW7:15.0Finished the same job 87 times meanwhileProcessing · 7 min 15 s1,292.4 KRW7:20.0Finished the same job 88 times meanwhileProcessing · 7 min 20 s1,307.3 KRW7:25.0Finished the same job 89 times meanwhileProcessing · 7 min 25 s1,322.1 KRW7:30.0Finished the same job 90 times meanwhileProcessing · 7 min 30 s1,337.0 KRW7:35.0Finished the same job 91 times meanwhileProcessing · 7 min 35 s1,351.9 KRW7:40.0Finished the same job 92 times meanwhileProcessing · 7 min 40 s1,366.7 KRW7:45.0Finished the same job 93 times meanwhileProcessing · 7 min 45 s1,381.6 KRW7:50.0Finished the same job 94 times meanwhileProcessing · 7 min 50 s1,396.4 KRW7:55.0Finished the same job 95 times meanwhileProcessing · 7 min 55 s1,411.3 KRWDone · 5.0 s0.65 KRWDone · 176.1 sLAS: 35 times meanwhile1,760.3 KRW7:57.8Finished the same job 95 times meanwhileDone · 477.8 sLAS: 95 times meanwhile1,419.6 KRWvs. Claude Fable 5.135× faster at 1/2,708 of the costvs. GPT-6 Astra95× faster at 1/2,184 of the cost

LTS measurement 2026-09-29 · three sets of primary-contractor order spreadsheets, carrier dispatches, DTG trip records · the same full input processed at once · median of 3 runs per engine · public API rates · exchange rate KRW 1,359.4 · all three engines correct in all 3 runs · video at 60× speed

01 Determination

LAS cross-checks order and dispatch tables and determineswhich primary contractor receives each dispatch's transport data

Order table · two rowsShipperPickupDrop-offPlate no.Shipper AShipper BState computed onceVehicle signals · valuesState builder describes each segmentTable type?Order listDispatch sheetOrder listDispatch sheetWhich column is pickup?Col 1Col 2Col 3Col 1Col 2Col 3Hazmat in first row?YesNoYesNoPlate number present?YesNoYesNoISO 14083 field coverage?12341234ChoiceTable typeShipment order listConfidence 1.00Yes/NoFirst-row hazmatProbability of true 0.12ScoreISO 14083 fields metExpected level 2.981234Probabilityof true01Low thresholdHigh thresholdConfirm No at onceLAS second passConfirm Yes at once
Which shipper?WaitingEvaluatingShipper A .96 · Other .04Shipper A .96 · Other .04Which vehicle?WaitingEvaluating12가3456 .91 · Other .0912가3456 .91 · Other .09Risky driving?WaitingEvaluatingHard braking .88 · None .12Hard braking .88 · None .12Mixed-load split?WaitingEvaluating3 loads .93 · 1 load .073 loads .93 · 1 load .07
{█"shipper": "Shipper A","vehicle": "12가3456","riskyDriving": "2 hard brakes","mixedLoad": true,"ship█"shipper": █"shipper": "Shipper A"█"shipper": "Shipper A",█"vehicle": █"vehicle": "12가3█"vehicle": "12가3456"█"vehicle": "12가3456",█"risky█"riskyDriving": █"riskyDriving": "2 hard█"riskyDriving": "2 hard brakes",█"mixed█"mixedLoad": █"mixedLoad": true█"mixedLoad": true,█"assign█"assignment": █"assignment": "do█"assignment": "done"█"assignment": "done"}
1 truck · 3 consolidated loads1 trip · one fuel signal per vehicleContractor A · 3.2 tSeoul→DaejeonContractor B · 2.6 tSeoul→DaeguContractor C · 1.8 tSeoul→BusanVehicle signal · LTS DTGFuel rate, distance, time — not split by cargoOrder · dispatch dataOrder · dispatch dataLAS engineMaps order × dispatch × vehicle signalWhich cargo was loaded?Which cargo was loaded?Do segment and time match?Do segment and time match?Tonne-km for each cargo?Tonne-km for each cargo?Tonne-km allocationTonne-km allocationDeterminingDeterminedContractor A cargo3.2 t × Seoul→Daejeon distance= tonne-km → cargo's emissionsContractor A cargo3.2 t × Seoul→Daejeon distance= tonne-km → cargo's emissionsContractor A viewScope 3 transport emissionsContractor B cargo2.6 t × Seoul→Daegu distance= tonne-km → cargo's emissionsContractor B cargo2.6 t × Seoul→Daegu distance= tonne-km → cargo's emissionsContractor B viewScope 3 transport emissionsContractor C cargo1.8 t × Seoul→Busan distance= tonne-km → cargo's emissionsContractor C cargo1.8 t × Seoul→Busan distance= tonne-km → cargo's emissionsContractor C viewScope 3 transport emissionsUnclassified cargo → LAS second-pass determinationAmbiguous cargo is checked against all candidate ordersin the LAS second pass and confirmed · this trip: 0

Time and cost to process the same full input at once

Even as the number of dispatches grows,LAS completes determination within seconds

DispatchesLAS engineClaude Fable 5.1GPT-6 Astra
92.16 s · KRW 0.2626.5 s · KRW 224.848.4 s · KRW 130.9
454.41 s · KRW 0.4963.2 s · KRW 569.2158.9 s · KRW 443.6
1805.03 s · KRW 0.65176.1 s · KRW 1,760.3477.8 s · KRW 1,419.6

LTS measurement 2026-09-29 · median of 3 runs per engine · public API rates (per 1 million input tokens) · exchange rate KRW 1,359.4

Determination accuracy

The LAS notation dictionary resolves 60.0% of target dispatches,while the full LAS engine resolves 97.2% of them

Talk to us

LTS presents a LAS determination setup that fitsthe primary contractor's order and dispatch formats

In a consultation, an LTS representative explains the scope and terms of adoption for your fleet size and partner-carrier structure.

Request a consultation

We use your details only to respond to your request.

LAS | LTS