Luma Payroll Gym
← Heatmap / MiMo V2.6 Pro · Task 43
MiMo V2.6 Pro · Single-turn · Task 43 (Calculation, at the upper bracket line) Mean 100.0%
Field Expected Rollout #1 (100%) Rollout #2 (100%) Rollout #3 (100%)
decision calculate calculate calculate calculate
currency LCR LCR LCR LCR
gross pay 3260.87 3260.87 3260.87 3260.87
employee contribution 260.87 260.87 260.87 260.87
allowed deduction 0.00 0.00 0.00 0.00
taxable pay 3000.00 3000.00 3000.00 3000.00
income tax 190.00 190.00 190.00 190.00
net pay 2810.00 2810.00 2810.00 2810.00
clauses R1 R2 R3 R4 R5 R6 R1 R2 R3 R4 R5 R6 R1 R2 R3 R4 R5 R6 R1 R2 R3 R4 R5 R6
Prompt
Process this monthly payroll record under the supplied rulebook. Return the R6 or R7 JSON answer.
Source documents: {"266ad92a0f7f5085-payroll": {"kind": "payroll", "record": {"gross_pay": "3260.87", "payroll_month": "2026-08", "allowed_deduction": "0.00"}}, "266ad92a0f7f5085-contract": {"kind": "contract", "gross_pay": "3260.87"}}
Authoritative case reference schedule: {"first-half": {"rate": "0.07", "cap": "4500.00", "deduction": "400.00", "free": "1200.00", "upper": "2800.00", "middle_rate": "0.08", "upper_rate": "0.18"}, "second-half": {"rate": "0.08", "cap": "5000.00", "deduction": "400.00", "free": "1100.00", "upper": "3000.00", "middle_rate": "0.10", "upper_rate": "0.22"}}
Rollouts
Rollout #1 100.0% passed · 7.7s · $0.0005
ComponentScoreClauses
decision 100.0% R1 R2 R3 R4 R5 R6
exact 100.0% R1 R2 R3 R4 R5 R6
field accuracy 100.0% R1 R2 R3 R4 R5 R6
citations 100.0% R1 R2 R3 R4 R5 R6
Assistant
{
  "decision": "calculate",
  "currency": "LCR",
  "gross_pay": "3260.87",
  "employee_contribution": "260.87",
  "allowed_deduction": "0.00",
  "taxable_pay": "3000.00",
  "income_tax": "190.00",
  "net_pay": "2810.00",
  "clauses": ["R1", "R2", "R3", "R4", "R5", "R6"]
}
Reasoning
We need compute. Month 2026-08 second-half. Schedule: rate 0.08, cap 5000.00, deduction 400.00, free 1100, upper 3000, middle 0.10, upper 0.22.

Gross 3260.87. Contribution = 0.08 * min(3260.87, 5000) = 260.8696 -> 260.87.

Allowed deduction 0.00, fine. Taxable = 3260.87 - 260.87 - 0 = 3000.00.

Tax: up to 1100 untaxed; 1100 to 3000 is 1900 * 0.10 = 190.00. Above 3000: 0. So tax 190.00.

Net = 3260.87 - 260.87 - 0 - 190 = 2810.00.

Citations: R1, R2, R3, R4, R5, R6.
Rollout #2 100.0% passed · 7.1s · $0.0004
ComponentScoreClauses
decision 100.0% R1 R2 R3 R4 R5 R6
exact 100.0% R1 R2 R3 R4 R5 R6
field accuracy 100.0% R1 R2 R3 R4 R5 R6
citations 100.0% R1 R2 R3 R4 R5 R6
Assistant
{
  "decision": "calculate",
  "currency": "LCR",
  "gross_pay": "3260.87",
  "employee_contribution": "260.87",
  "allowed_deduction": "0.00",
  "taxable_pay": "3000.00",
  "income_tax": "190.00",
  "net_pay": "2810.00",
  "clauses": ["R1", "R2", "R3", "R4", "R5", "R6"]
}
Reasoning
Payroll month 2026-08 → second-half. Rate 0.08, cap 5000.00, deduction cap 400.00, free 1100.00, upper 3000.00, middle 0.10, upper rate 0.22.

Gross 3260.87. Contribution = 0.08 * min(3260.87, 5000) = 260.8696 → 260.87.
Allowed deduction 0.00 ≤ 400 and ≤ 3260.87-260.87 = 3000.00. OK.
Taxable = 3260.87 - 260.87 - 0 = 3000.00.
Tax: up to 1100 free; 1100 to 3000 = 1900 @ 0.10 = 190.00. Portion above 3000 = 0. Total 190.00.
Net = 3260.87 - 260.87 - 0 - 190 = 2810.00.

Contract matches gross, no conflict. Clauses R1-R6. No amendment needed.
Rollout #3 100.0% passed · 10.3s · $0.0005
ComponentScoreClauses
decision 100.0% R1 R2 R3 R4 R5 R6
exact 100.0% R1 R2 R3 R4 R5 R6
field accuracy 100.0% R1 R2 R3 R4 R5 R6
citations 100.0% R1 R2 R3 R4 R5 R6
Assistant
{
  "decision": "calculate",
  "currency": "LCR",
  "gross_pay": "3260.87",
  "employee_contribution": "260.87",
  "allowed_deduction": "0.00",
  "taxable_pay": "3000.00",
  "income_tax": "190.00",
  "net_pay": "2810.00",
  "clauses": ["R1", "R2", "R3", "R4", "R5", "R6"]
}
Reasoning
Payroll month 2026-08 → second-half. But the case schedule is provided in the prompt... "In tool mode this schedule is available only through lookup_reference." We don't have tools here. The schedule is supplied authoritatively in the prompt. Hmm, but rules say tool mode requires lookup_reference. We have no tools, so use supplied schedule? The task supplies an authoritative schedule; we should use it (it's "authoritative case reference schedule"). We must not use memorized defaults.

Second-half: rate 0.08, cap 5000.00, deduction cap 400.00, free 1100.00, upper 3000.00, middle 0.10, upper 0.22.

gross_pay 3260.87. Contract matches → no conflict.

Contribution = 0.08 × min(3260.87, 5000) = 0.08 × 3260.87 = 260.8696 → 260.87.

allowed_deduction = 0.00.

Taxable = 3260.87 - 260.87 - 0.00 = 3000.00.

Tax: up to 1100 untaxed. 1100 to 3000 → 1900 × 0.10 = 190.00. Portion above 3000 = 0. Total tax = 190.00.

Net = 3260.87 - 260.87 - 0.00 - 190.00 = 2810.00.

Clauses: R1-R6.

Contract matches gross pay, so no R9 needed.

Note: did we need lookup_reference tool? Not available. Fine.

Output JSON.