alea-institute/alea-r3-moe-2b-a417m
Text Generation • 2B • Updated • 133
Datasets:
id stringlengths 17 20 | problem stringlengths 73 459 | steps listlengths 1 4 | answer stringlengths 1 95 | final stringlengths 14 123 | program stringlengths 11 874 | executed_output stringlengths 1 95 | type stringclasses 15
values | subtype stringclasses 108
values | setting stringclasses 278
values | origin stringclasses 3
values | steps_recomputed int64 0 4 | generator stringclasses 1
value | verification stringclasses 1
value | license stringclasses 1
value |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
wbab7aaeb358439_0 | To prepare for a road trip, Ludger and Beniamin are scaling a trail-mix recipe that serves 5 people to serve 8. The recipe calls for 275 grams of dried fruit. How many grams of dried fruit should they pack? | [
"8 ÷ 5 = 1.6",
"275 × 1.6 = 440 grams",
"440 grams of dried fruit"
] | 440 grams | They should pack 440 grams of dried fruit. | answer = 275 * (8 / 5) | 440.0 | rates | scaling a recipe | a road trip | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wbab7aaeb358439_1 | With laundry day coming up, Hermina and Michalina have a $27.80 wash-and-fold bill. The bill includes 3 shirts at $4.35 each and a $14.75 wash-and-fold charge. They want to leave a 16% tip on the bill before the tip. How much should they leave as a tip? | [
"3 × $4.35 = $13.05",
"$13.05 + $14.75 = $27.80",
"$27.80 × 0.16 = $4.448",
"$4.448 rounded to the nearest cent = $4.45"
] | $4.45 | They should leave a $4.45 tip. | answer = round((3 * 4.35 + 14.75) * 0.16, 2) | 4.45 | percent | a tip on a bill | a laundry schedule | batch-1 | 3 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wbab7aaeb358439_3 | To finish picking apples at the orchard, Edme and Pierrette start at 9:38 AM and work for 2 hours and 47 minutes. What time do they finish? | [
"9:38 AM + 2 hours = 11:38 AM",
"11:38 AM + 47 minutes = 12:25 PM"
] | 12:25 PM | They finish at 12:25 PM. | from datetime import date, datetime, timedelta
start = datetime.strptime('9:38 AM', '%I:%M %p')
answer = (start + timedelta(hours=2, minutes=47)).strftime('%-I:%M %p') | 12:25 PM | clock | add a duration to a start time | an orchard | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wbab7aaeb358439_4 | A swimmer at the pool plans to swim 1.35 miles at a steady pace of 4.6 minutes per mile. How long will the swim take? | [
"1.35 × 4.6 = 6.21 minutes",
"6.21 minutes = 6 minutes and 0.21 × 60 = 12.6 seconds"
] | 6.21 minutes | The swim will take 6.21 minutes. | answer = 1.35 * 4.6 | 6.21 | durations | time needed at a given pace (minutes per mile or page) | a swimming pool | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wbab7aaeb358439_5 | Every theater rehearsal, Masahiko and Veijo record how many minutes they spend practicing their lines. Their times over four rehearsals are 38, 45, 41, and 52 minutes. What is their mean practice time? | [
"38 + 45 = 83 minutes",
"41 + 52 = 93 minutes",
"83 + 93 = 176 minutes",
"176 ÷ 4 = 44 minutes"
] | 44 minutes | Their mean practice time is 44 minutes. | answer = (38 + 45 + 41 + 52) / 4 | 44.0 | averages | mean of several values | a theater rehearsal | batch-1 | 4 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wbab7aaeb358439_6 | During a city bus route training course, Brahim and Denisa need an average score of 82 across five quizzes. Their first four scores are 76, 84, 79, and 88. What score must they earn on the fifth quiz to reach the target average? | [
"82 × 5 = 410 points needed in total",
"76 + 84 = 160 points",
"79 + 88 = 167 points; 160 + 167 = 327 points so far",
"410 − 327 = 83 points"
] | 83 points | They must earn 83 points on the fifth quiz. | answer = 82 * 5 - (76 + 84 + 79 + 88) | 83 | averages | the missing score needed for a target average | a city bus route | batch-1 | 3 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wbab7aaeb358439_7 | After checking a map before the ferry crossing, Annemie and Matylda see that the route measures 8.6 centimeters on the map. The map scale is 1 centimeter to 2.7 kilometers. How long is the actual crossing route in kilometers? | [
"8.6 × 2.7 = 23.22 kilometers",
"23.22 kilometers × 1,000 = 23,220 meters",
"23,220 meters ÷ 1,000 = 23.22 kilometers"
] | 23.22 kilometers | The actual crossing route is 23.22 kilometers long. | answer = 8.6 * 2.7 | 23.22 | rates | a map scale | a ferry crossing | batch-1 | 3 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
we1c1e106596d86_0 | Genowefa’s phone plan cost $11.80 per week for four weeks, but the provider raised the weekly price to $14.16. What was the percent increase in the four-week cost? | [
"Old four-week cost: $11.80 × 4 = $47.20",
"New four-week cost: $14.16 × 4 = $56.64",
"Increase: $56.64 − $47.20 = $9.44",
"Percent increase: ($9.44 ÷ $47.20) × 100 = 20%"
] | 20% | The four-week phone plan cost increased by 20%. | old_cost = 11.80 * 4
new_cost = 14.16 * 4
answer = round((new_cost - old_cost) / old_cost * 100, 2) | 20.0 | percent | a percent increase | a phone plan | batch-1 | 4 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
we1c1e106596d86_1 | At the community hall, Neeltje’s recital begins at 6:38 PM and lasts 1 hour 47 minutes. When will it end? | [
"1 hour = 60 minutes",
"Duration: 60 + 47 = 107 minutes",
"6:38 PM + 107 minutes = 8:25 PM"
] | 8:25 PM | The recital will end at 8:25 PM. | from datetime import datetime, timedelta
time = datetime.strptime('6:38 PM', '%I:%M %p') + timedelta(minutes=107)
answer = time.strftime('%I:%M %p').lstrip('0') | 8:25 PM | clock | add a duration to a start time | a recital | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
we1c1e106596d86_2 | A court filing deadline falls on May 7, 2025. Evren was born on August 19, 1998. How old will Evren be on the deadline? | [
"2025 − 1998 = 27 years",
"August 19 falls after May 7, so the 2025 birthday has not occurred",
"27 − 1 = 26 years"
] | 26 years old | Evren will be 26 years old on the filing deadline. | from datetime import date
deadline = date(2025, 5, 7)
birthday = date(1998, 8, 19)
answer = deadline.year - birthday.year - ((deadline.month, deadline.day) < (birthday.month, birthday.day)) | 26 | dates | age on a given date from a birthday | a court filing deadline | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
we1c1e106596d86_3 | During a shift at Issa’s toy store in Denver, Rolf checks the time in Boston, where the clock is two hours ahead. If it is 2:35 PM in Denver, what time is it in Boston? | [
"Two hours × 60 = 120 minutes",
"2:35 PM + 120 minutes = 4:35 PM"
] | 4:35 PM | It is 4:35 PM in Boston. | from datetime import datetime, timedelta
time = datetime.strptime('2:35 PM', '%I:%M %p') + timedelta(minutes=2 * 60)
answer = time.strftime('%I:%M %p').lstrip('0') | 4:35 PM | clock | a fixed time-zone offset | a toy store | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
we1c1e106596d86_4 | With a big cleanup underway, Jostein’s crew needs 24 bottles of cleaner. A bulk deal offers 12 bottles for $28.80, while single bottles cost $2.65 each; how much does the crew save by choosing the bulk deal? | [
"Savings: (24 × $2.65) − (2 × $28.80) = $6.00"
] | $6.00 | The crew saves $6.00 by choosing the bulk deal. | answer = round(24 * 2.65 - 2 * 28.80, 2) | 6.0 | money | a bulk deal versus singles | a cleaning crew | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
we1c1e106596d86_6 | $286.75 was set aside for Bahareh and Nikodem’s woodworking project, which uses 40 boards. If 15% of the boards need sanding, how many boards need sanding? | [
"40 × 0.15 = 6",
"6 boards need sanding"
] | 6 boards | Six boards need sanding. | answer = 40 * 0.15 | 6.0 | percent | a percent of a group | a woodworking project | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
we1c1e106596d86_7 | To meet a court filing deadline, Noriyuki uses a filing service that charges $84.50 before a 20% discount and 6% sales tax. How much does he pay after the discount and tax? | [
"Discount: $84.50 × 0.20 = $16.90",
"Price after discount: $84.50 − $16.90 = $67.60",
"Sales tax: $67.60 × 0.06 = $4.056",
"Total: $67.60 + $4.056 = $71.656 ≈ $71.66"
] | $71.66 | Noriyuki pays $71.66 after the discount and sales tax. | answer = round((84.50 * (1 - 0.20)) * (1 + 0.06) + 1e-9, 2) | 71.66 | percent | a discount followed by sales tax | a court filing deadline | batch-1 | 4 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wd22415790bf508_0 | A dentist office has Eckhard schedule 5 patient cleanings that take 34, 41, 38, 46, and 36 minutes. What is the average time per cleaning? | [
"34 + 41 + 38 + 46 + 36 = 195 minutes",
"195 ÷ 5 = 39 minutes per cleaning"
] | 39 minutes | The average cleaning takes 39 minutes. | answer = (34 + 41 + 38 + 46 + 36) / 5 | 39.0 | durations | average time per task | a dentist office | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wd22415790bf508_1 | This canoe rental shop collected $162 from Romuald for a rental and $675 in total that day. What percent of the day's total did Romuald pay? | [
"162 ÷ 675 = 0.24",
"0.24 × 100 = 24%"
] | 24% | Romuald paid 24% of the day's total. | answer = 162 / 675 * 100 | 24.0 | percent | what percent one amount is of another | a canoe rental | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wd22415790bf508_3 | The piano lesson group has 25 students, and 40% of them, including Luana, are learning a new piece this week. How many students are learning it? | [
"25 × 0.40 = 10 students"
] | 10 students | 10 students are learning the new piece this week. | answer = 25 * 0.40 | 10.0 | percent | a percent of a group | a piano lesson | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wd22415790bf508_4 | The marathon route is 42 kilometers long, and Masayoshi finishes in 3 hours and 27 minutes. Takanori wants to know Masayoshi's average speed in kilometers per hour. What was it? | [
"3 × 60 = 180 minutes",
"180 + 27 = 207 minutes; 207 ÷ 60 = 3.45 hours",
"42 ÷ 3.45 = 12.1739130435 kilometers per hour"
] | 12.1739130435 kilometers per hour | Masayoshi's average speed was about 12.17 kilometers per hour. | answer = 42 / (3 + 27 / 60) | 12.17391304347826 | rates | a unit rate | a marathon | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wd22415790bf508_5 | Thirty-five dollars is the sale price Betina paid at a parking garage after a 30% discount. What was the original price before the discount? | [
"100% − 30% = 70% of the original price",
"$35.00 ÷ 0.70 = $50.00"
] | $50.00 | The original price was $50.00. | answer = round(35 / (1 - 0.30), 2) | 50.0 | percent | the original price from a sale price and discount | a parking garage | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wd22415790bf508_6 | The meal prep plan's timer is set for 18 minutes to chop vegetables, 27 minutes to roast them, and 8 minutes to let them cool. Linn starts the timer at 5:12 PM, and Agnete wants to know when the plan will be finished. What time will it be? | [
"18 + 27 = 45 minutes",
"45 + 8 = 53 minutes",
"5:12 PM + 53 minutes = 6:05 PM"
] | 6:05 PM | The meal prep plan will be finished at 6:05 PM. | from datetime import date, datetime, timedelta
answer = (datetime.strptime('5:12 PM', '%I:%M %p') + timedelta(minutes=18 + 27 + 8)).strftime('%I:%M %p').lstrip('0') | 6:05 PM | clock | a timer or alarm | a meal prep plan | batch-1 | 3 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wd22415790bf508_7 | After a two-night hotel stay, Mattias and Walerian receive a bill with room charges of $348.60 and a $27.40 service charge. They want to leave a 18% tip on the combined bill before any tip is added. How much should the tip be? | [
"$348.60 + $27.40 = $376.00",
"$376.00 × 0.18 = $67.68",
"Tip amount = $67.68"
] | $67.68 | They should leave a tip of $67.68. | answer = round((348.60 + 27.40) * 0.18, 2) | 67.68 | percent | a tip on a bill | a hotel stay | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wdc7f00362c7777_1 | Because Svend, a pet groomer, spent 38 minutes on each of 4 grooming sessions and 52 minutes on each of 3 more sessions this week, he wants to total his grooming time. How much time did he spend grooming altogether? | [
"38 × 4 = 152 minutes; 52 × 3 = 156 minutes",
"152 + 156 = 308 minutes"
] | 308 minutes | Svend spent 308 minutes grooming this week. | answer = 38 * 4 + 52 * 3 | 308 | durations | total time over several sessions in a week | a pet groomer | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wdc7f00362c7777_2 | $2,480 was the public library's monthly printing budget before it was reduced by 12.5%. How much is the budget after the decrease? | [
"$2,480 × 0.125 = $310",
"$2,480 − $310 = $2,170"
] | $2,170 | The library's budget after the decrease is $2,170. | answer = round(2480 - 2480 * 0.125, 2) | 2170.0 | percent | a percent decrease | a public library | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wdc7f00362c7777_3 | To settle a small office's $486.75 supply bill, Jalmari applies a $38.60 credit for returned items. What amount does the office still owe? | [
"$486.75 − $38.60 = $448.15",
"Amount owed = $448.15"
] | $448.15 | The office still owes $448.15. | answer = round(486.75 - 38.60, 2) | 448.15 | money | a refund or credit applied to a bill | a small office | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wdc7f00362c7777_4 | At the cooking class, Martynas combines 1.35 liters of broth with 680 milliliters of water, then adds 425 milliliters of cream. Laureano pours the mixture equally into 5 containers. How much mixture goes into each container, in milliliters? | [
"1.35 L × 1,000 = 1,350 mL",
"1,350 + 680 = 2,030 mL",
"2,030 + 425 = 2,455 mL",
"2,455 ÷ 5 = 491 mL"
] | 491 milliliters | Each container gets 491 milliliters of mixture. | answer = (1.35 * 1000 + 680 + 425) / 5 | 491.0 | units | liters and milliliters | a cooking class | batch-1 | 4 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wdc7f00362c7777_5 | When Jarno and Farida tracked sales at a holiday market over four shifts, they sold 37, 46, 52, and 41 ornaments. What was their average number of ornaments sold per shift? | [
"37 + 46 = 83",
"83 + 52 + 41 = 176",
"176 ÷ 4 = 44 ornaments per shift"
] | 44 ornaments per shift | They sold an average of 44 ornaments per shift. | answer = (37 + 46 + 52 + 41) / 4 | 44.0 | averages | mean of several values | a holiday market | batch-1 | 3 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wdc7f00362c7777_6 | Kazue is preparing a snack recipe for a ferry crossing. The recipe serves 6 people and uses 225 grams of oats, 180 grams of dried fruit, and 90 grams of seeds; she needs enough for 10 people. How many grams of each ingredient should she use? | [
"10 ÷ 6 = 5/3 scaling factor",
"225 × 5/3 = 375 g of oats",
"180 × 5/3 = 300 g of dried fruit",
"90 × 5/3 = 150 g of seeds"
] | 375 g of oats, 300 g of dried fruit, and 150 g of seeds | Kazue should use 375 g of oats, 300 g of dried fruit, and 150 g of seeds. | factor = 10 / 6
answer = f"{225 * factor:.0f} g of oats, {180 * factor:.0f} g of dried fruit, and {90 * factor:.0f} g of seeds" | 375 g of oats, 300 g of dried fruit, and 150 g of seeds | rates | scaling a recipe | a ferry crossing | batch-1 | 3 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wdc7f00362c7777_7 | When Shai has a 2-hour-24-minute tutoring session, Emirhan divides the time equally into 6 study blocks. How long is each block? | [
"2 hours 24 minutes = 144 minutes; 144 ÷ 6 = 24 minutes"
] | 24 minutes | Each study block is 24 minutes long. | answer = (2 * 60 + 24) // 6 | 24 | durations | split a time span into equal parts | a tutoring center | batch-1 | 0 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wa3821b7a07066b_0 | For the pool’s first Saturday session, Torstein and Caterina arrived at 10:18 AM and left at 1:07 PM. How long were they at the pool? | [
"10:18 AM to 12:00 PM = 1 hour 42 minutes",
"12:00 PM to 1:07 PM = 1 hour 7 minutes",
"1 hour 42 minutes + 1 hour 7 minutes = 2 hours 49 minutes"
] | 2 hours 49 minutes | They were at the pool for 2 hours 49 minutes. | from datetime import datetime
start = datetime.strptime('10:18 AM', '%I:%M %p')
end = datetime.strptime('1:07 PM', '%I:%M %p')
answer = f'{int((end-start).total_seconds()//3600)} hours {int((end-start).total_seconds()//60%60)} minutes' | 2 hours 49 minutes | clock | a time span that crosses noon | a swimming pool | batch-1 | 0 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wa3821b7a07066b_1 | The community garden’s seedling sale is 5 weeks and 4 days from today. Blandine and Kimiko are making signs for it. How many days away is the sale? | [
"5 × 7 = 35 days",
"35 + 4 = 39 days"
] | 39 days | The seedling sale is 39 days away. | answer = 5 * 7 + 4 | 39 | dates | weeks and days until an event | a community garden | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wa3821b7a07066b_3 | Each pottery class project is scored out of 100 points. Leonhard has scores of 78, 85, and 91, and needs an average of 86 over four projects; what score does he need on the last project? | [
"4 × 86 − (78 + 85 + 91) = 90"
] | 90 points | Leonhard needs 90 points on his last project. | answer = 4 * 86 - (78 + 85 + 91) | 90 | averages | the missing score needed for a target average | a pottery class | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wa3821b7a07066b_4 | Since the city bus route ends in a time zone 2 hours ahead of its starting city, Lindita and Antonie check the schedule for a bus leaving at 9:35 AM and traveling for 3 hours 48 minutes. What local time will it arrive? | [
"9:35 AM + 3 hours 48 minutes = 1:23 PM",
"1:23 PM + 2 hours = 3:23 PM",
"Arrival in the destination time zone = 3:23 PM"
] | 3:23 PM | The bus will arrive at 3:23 PM local time. | from datetime import datetime, timedelta
depart = datetime.strptime('9:35 AM', '%I:%M %p')
arrival = depart + timedelta(hours=3, minutes=48) + timedelta(hours=2)
answer = arrival.strftime('%I:%M %p').lstrip('0') | 3:23 PM | clock | a fixed time-zone offset | a city bus route | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wa3821b7a07066b_7 | During a lemonade stand shift, Gervase and Adina want to make 2.4 times their recipe, which uses 375 milliliters of lemon juice. How much lemon juice will they need? | [
"375 × 2.4 = 900 milliliters"
] | 900 milliliters | They will need 900 milliliters of lemon juice. | answer = 375 * 2.4 | 900.0 | rates | scaling a recipe | a lemonade stand | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w76df593d83087f_0 | At Kjersti and Agim’s orchard, a basket of pears is marked $37.60. It is discounted by 15%, and 6.5% sales tax is added to the discounted price. What is the total price? | [
"$37.60 × 0.85 = $31.96",
"$31.96 × 1.065 = $34.0374, which rounds to $34.04"
] | $34.04 | The total price is $34.04. | answer = round(37.60 * (1 - 0.15) * (1 + 0.065), 2) | 34.04 | percent | a discount followed by sales tax | an orchard | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w76df593d83087f_1 | During leap-year February, Lianne parks at a garage for every day from February 1 through February 29. The garage charges $8.75 per day. How much does she pay? | [
"February 1 through February 29 = 29 days",
"29 × $8.75 = $253.75"
] | $253.75 | Lianne pays $253.75. | answer = 29 * 8.75 | 253.75 | dates | a February in a leap year | a parking garage | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w76df593d83087f_4 | $36.50 is Kornelis’s hourly pay for his shift processing court filings before a deadline. He works from 8:20 AM to 4:50 PM and takes an unpaid 35-minute break. How many hours does he get paid for? | [
"8:20 AM to 4:50 PM = 8 hours 30 minutes",
"8 hours 30 minutes − 35 minutes = 7 hours 55 minutes"
] | 7 hours 55 minutes | Kornelis gets paid for 7 hours 55 minutes. | answer = 8 * 60 + 30 - 35 | 475 | clock | a shift with an unpaid break | a court filing deadline | batch-1 | 0 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w76df593d83087f_5 | A hotel has 86 bath towels in storage at the start of Francesc and Rochus’s stay. On Monday, guests use 23 towels, and the hotel receives 17 clean ones. On Tuesday, guests use 19 towels. How many towels remain? | [
"86 − 23 = 63 towels",
"63 + 17 = 80 towels",
"80 − 19 = 61 towels"
] | 61 towels | There are 61 towels remaining. | answer = 86 - 23 + 17 - 19 | 61 | multistep | inventory changing over several days | a hotel stay | batch-1 | 3 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w76df593d83087f_6 | Last, Dorit’s dog-walking business sets a timer for a walk that starts at 2:18 PM. The walk lasts 47 minutes, and Miodrag then spends 16 minutes returning the dogs and 9 minutes writing the visit notes. At what time is all the work finished? | [
"2:18 PM + 47 minutes = 3:05 PM",
"3:05 PM + 16 minutes = 3:21 PM",
"3:21 PM + 9 minutes = 3:30 PM",
"The work is finished at 3:30 PM"
] | 3:30 PM | All the work is finished at 3:30 PM. | from datetime import datetime, timedelta
answer = (datetime.strptime('2:18 PM', '%I:%M %p') + timedelta(minutes=47 + 16 + 9)).strftime('%-I:%M %p') | 3:30 PM | clock | a timer or alarm | a dog walking business | batch-1 | 3 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wb9f80c93ce263b_0 | This dog-walking business is comparing two leash-bag packages. Svante found 18 bags for $14.58, while Chiharu found 25 bags for $19.25. Which package has the lower price per bag, and by how much? | [
"$14.58 ÷ 18 = $0.81 per bag",
"$19.25 ÷ 25 = $0.77 per bag",
"$0.81 − $0.77 = $0.04 per bag"
] | The 25-bag package, by $0.04 per bag | The 25-bag package has the lower price, by $0.04 per bag. | price_a = 14.58 / 18
price_b = 19.25 / 25
answer = round(price_a - price_b, 2) | 0.04 | money | comparing unit prices | a dog walking business | batch-1 | 3 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wb9f80c93ce263b_1 | After a quilting circle meets on Tuesday, Brunon counts 19 days until its next meeting. What weekday will the circle meet? | [
"Tuesday + 19 days = Sunday"
] | Sunday | The quilting circle will meet on Sunday. | from datetime import date, timedelta
start = date(2024, 1, 2) # Tuesday
answer = (start + timedelta(days=19)).strftime('%A') | Sunday | dates | the weekday N days later | a quilting circle | batch-1 | 0 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wb9f80c93ce263b_2 | Before applying a credit, Boubacar's tailor shop bill for Kiyoshi is $186.75. The shop refunds $28.40 for a hem adjustment, then applies a $16.85 credit for a delayed fitting. What amount does Kiyoshi owe? | [
"$186.75 − $28.40 = $158.35",
"$158.35 − $16.85 = $141.50",
"Amount owed = $141.50"
] | $141.50 | Kiyoshi owes $141.50. | answer = round(186.75 - 28.40 - 16.85, 2) | 141.5 | money | a refund or credit applied to a bill | a tailor shop | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wb9f80c93ce263b_3 | While helping at a wedding reception, Uberto and Vitalis keep score in a table quiz. The team has scored 76, 83, and 91 points in its first three rounds. What score does it need in the fourth round to average 85 points across all four rounds? | [
"85 × 4 = 340 points needed total",
"76 + 83 + 91 = 250 points so far",
"340 − 250 = 90 points needed",
"Required fourth-round score = 90 points"
] | 90 points | The team needs 90 points in the fourth round to average 85. | answer = 85 * 4 - (76 + 83 + 91) | 90 | averages | the missing score needed for a target average | a wedding reception | batch-1 | 3 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wb9f80c93ce263b_4 | For a long weekend, Eligio is waiting at an airport during a flight connection and tracks snack-box inventory at the transfer lounge. The lounge starts with 58 boxes; on Friday, 21 are sold and 34 arrive, on Saturday 26 are sold, and on Sunday 18 arrive while 23 are sold. How many boxes remain after Sunday? | [
"58 − 21 + 34 = 71 boxes",
"71 − 26 = 45 boxes",
"45 + 18 − 23 = 40 boxes"
] | 40 boxes | There are 40 snack boxes left after Sunday. | answer = 58 - 21 + 34 - 26 + 18 - 23 | 40 | multistep | inventory changing over several days | a flight connection | batch-1 | 3 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wb9f80c93ce263b_6 | Since the cleaning crew starts at 10:38 AM and finishes at 1:17 PM, Masahide and Antonija work continuously without a break. How long is their shift? | [
"10:38 AM to 1:17 PM = 2 hours 39 minutes"
] | 2 hours 39 minutes | Their shift lasts 2 hours 39 minutes. | from datetime import datetime
start = datetime.strptime('10:38 AM', '%I:%M %p')
end = datetime.strptime('1:17 PM', '%I:%M %p')
minutes = int((end - start).total_seconds() // 60)
answer = f'{minutes // 60} hours {minutes % 60} minutes' | 2 hours 39 minutes | clock | a time span that crosses noon | a cleaning crew | batch-1 | 0 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wb9f80c93ce263b_7 | At the Riverside Park route, Muharrem and Leanid log three dog-walking sessions during the week: 48 minutes on Monday, 67 minutes on Wednesday, and 54 minutes on Saturday. How much total time do they spend walking dogs that week? | [
"48 + 67 = 115 minutes",
"115 + 54 = 169 minutes",
"169 minutes = 2 hours 49 minutes"
] | 169 minutes | They spend 169 minutes walking dogs that week. | answer = 48 + 67 + 54 | 169 | durations | total time over several sessions in a week | a dog walking business | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w499fa460f4f296_0 | While camping, Aurelie and Hidehiko take turns staffing the campsite’s welcome booth from 8:10 AM until 3:40 PM. Aurelie takes an unpaid 35-minute break during the shift. How many hours does she work, excluding her break? | [
"3:40 PM − 8:10 AM = 7 hours 30 minutes",
"7 hours 30 minutes − 35 minutes = 6 hours 55 minutes",
"6 hours 55 minutes = 415 minutes"
] | 6 hours 55 minutes | Aurelie works 6 hours 55 minutes, excluding her break. | answer = 7 * 60 + 30 - 35 | 415 | clock | a shift with an unpaid break | a camping trip | batch-1 | 0 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w499fa460f4f296_1 | Because the car rental lot is busy, Achiel and Iryna work together to clean returned cars. Achiel can clean 5 cars per hour, and Iryna can clean 4 cars per hour. How many cars can they clean together in 3 hours? | [
"5 + 4 = 9 cars per hour",
"9 × 3 = 27 cars"
] | 27 cars | Together, Achiel and Iryna can clean 27 cars in 3 hours. | answer = (5 + 4) * 3 | 27 | rates | two people working together | a car rental | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w499fa460f4f296_2 | When Lodewijk arrives at the craft fair, Natsumi tells him the pottery demonstration starts in 1 hour 35 minutes. If it is 10:25 AM now, what time does the demonstration start? | [
"10:25 AM + 1 hour = 11:25 AM",
"11:25 AM + 35 minutes = 12:00 PM"
] | 12:00 PM | The pottery demonstration starts at 12:00 PM. | from datetime import datetime, timedelta
answer = (datetime.strptime('10:25 AM', '%I:%M %p') + timedelta(hours=1, minutes=35)).strftime('%I:%M %p') | 12:00 PM | clock | add a duration to a start time | a craft fair | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w499fa460f4f296_3 | Muhamed buys Conny a homework planner marked at $28.65, with a 20% discount. The sales tax is 6% of the discounted price. What is the total cost after the discount and tax? | [
"$28.65 × 0.20 = $5.73",
"$28.65 − $5.73 = $22.92",
"$22.92 × 0.06 = $1.3752",
"$22.92 + $1.3752 = $24.2952 ≈ $24.30"
] | $24.30 | The homework planner costs $24.30 after the discount and tax. | answer = round((28.65 * (1 - 0.20)) * (1 + 0.06), 2) | 24.3 | percent | a discount followed by sales tax | a homework schedule | batch-1 | 4 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w499fa460f4f296_4 | On Tuesday, Grzegorz and Isa spend 2 hours 47 minutes choosing fabric, 1 hour 38 minutes comparing patterns, and 3 hours 26 minutes cutting samples at the fabric store. How much time do they spend on these tasks altogether? | [
"2 hours 47 minutes + 1 hour 38 minutes + 3 hours 26 minutes = 7 hours 51 minutes"
] | 7 hours 51 minutes | They spend 7 hours 51 minutes on the tasks altogether. | answer = (2 * 60 + 47) + (1 * 60 + 38) + (3 * 60 + 26) | 471 | durations | add several durations in hours and minutes | a fabric store | batch-1 | 0 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w499fa460f4f296_5 | To catch the county fair’s evening parade, Mladen and Teodor need to finish setting up their booth by 6:15 PM. Their setup takes 2 hours 35 minutes. What time should they start? | [
"6:15 PM − 2 hours = 4:15 PM",
"4:15 PM − 35 minutes = 3:40 PM",
"Start time = 3:40 PM"
] | 3:40 PM | They should start setting up at 3:40 PM. | from datetime import datetime, timedelta
answer = (datetime.strptime('6:15 PM', '%I:%M %p') - timedelta(hours=2, minutes=35)).strftime('%I:%M %p') | 03:40 PM | clock | work back from an end time to a start time | a county fair | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w499fa460f4f296_6 | Before running a marathon, Shinsuke puts $1,250 into an account earning simple interest at 3.6% per year. He leaves the money there for 8 months. How much interest does it earn? | [
"8 ÷ 12 = 2/3 year",
"$1,250 × 0.036 = $45 per year",
"$45 × 2/3 = $30"
] | $30.00 | The account earns $30.00 in simple interest. | answer = round(1250 * 0.036 * (8 / 12), 2) | 30.0 | percent | simple interest for some months or years | a marathon | batch-1 | 3 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w52f303bda81632_0 | To plan a solar panel install, Dionizy and Atilio set aside $8,460. They spend $2,735 on panels, $1,680 on mounting hardware, and $945 on permits and delivery. How much of the budget is left? | [
"$2,735 + $1,680 = $4,415",
"$4,415 + $945 = $5,360",
"$8,460 − $5,360 = $3,100"
] | $3,100 | They have $3,100 left in the budget. | answer = round(8460 - (2735 + 1680 + 945), 2) | 3100 | money | budget left after several purchases | a solar panel install | batch-1 | 3 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w52f303bda81632_1 | For the book club's Saturday meetup, Gudmund leaves at 1:18 PM and travels 37 minutes to Ilsa's home. After a 12-minute stop, they drive another 46 minutes to the bookstore. What time do they arrive? | [
"1:18 PM + 37 minutes = 1:55 PM",
"1:55 PM + 12 minutes = 2:07 PM",
"2:07 PM + 46 minutes = 2:53 PM"
] | 2:53 PM | They arrive at 2:53 PM. | from datetime import datetime, timedelta
start = datetime.strptime('1:18 PM', '%I:%M %p')
arrival = start + timedelta(minutes=37 + 12 + 46)
answer = arrival.strftime('%I:%M %p').lstrip('0') | 2:53 PM | multistep | a trip with legs, stops and arrival time | a book club | batch-1 | 3 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w52f303bda81632_2 | Last month, Anssi and Prince's pottery class made 84 glazed bowls. This month, they made 25% fewer bowls. How many bowls did they make this month? | [
"84 × 0.25 = 21 bowls",
"84 − 21 = 63 bowls"
] | 63 bowls | They made 63 bowls this month. | answer = 84 - 84 * 0.25 | 63.0 | percent | a percent decrease | a pottery class | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w52f303bda81632_6 | Next, Enric and Maarja finish a woodworking project at 5:12 PM. The final sanding took 48 minutes, and varnishing before that took 1 hour and 15 minutes. What time did they start varnishing? | [
"5:12 PM − 48 minutes = 4:24 PM",
"4:24 PM − 1 hour 15 minutes = 3:09 PM"
] | 3:09 PM | They started varnishing at 3:09 PM. | from datetime import datetime, timedelta
end = datetime.strptime('5:12 PM', '%I:%M %p')
start = end - timedelta(minutes=48 + 75)
answer = start.strftime('%I:%M %p').lstrip('0') | 3:09 PM | clock | work back from an end time to a start time | a woodworking project | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wce5ca8cd7aa1c3_0 | Yaacov starts his meal prep at 2:18 PM by chopping vegetables for 37 minutes, then immediately cooks them for 1 hour and 16 minutes. What time does he finish? | [
"37 + 76 = 113 minutes",
"2:18 PM + 113 minutes = 4:11 PM"
] | 4:11 PM | Yaacov finishes meal prep at 4:11 PM. | from datetime import datetime, timedelta
start = datetime.strptime("2:18 PM", "%I:%M %p")
answer = (start + timedelta(minutes=37 + 76)).strftime("%-I:%M %p") | 4:11 PM | clock | back-to-back activities with a schedule | a meal prep plan | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wce5ca8cd7aa1c3_1 | Before Semra closes her farmers market stall, she checks the day's sales: $186.40 came from vegetables, and total sales were $742.50. What percent of the sales came from vegetables? | [
"$186.40 ÷ $742.50 = 0.2511…",
"0.2511… × 100 = 25.11…",
"Round to the nearest tenth of a percent: 25.1%",
"Check: 25.1% of $742.50 ≈ $186.40"
] | 25.1% | About 25.1% of the day's sales came from vegetables. | answer = round(186.40 / 742.50 * 100, 1) | 25.1 | percent | what percent one amount is of another | a farmers market | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wce5ca8cd7aa1c3_2 | After reviewing a utility bill, Gorazd uses a service map to estimate the length of a buried water line. The map scale is 1:25,000, and the line measures 4.8 centimeters on the map. How long is the water line in kilometers? | [
"4.8 × 25,000 = 120,000 centimeters",
"120,000 ÷ 100,000 = 1.2 kilometers"
] | 1.2 kilometers | The water line is 1.2 kilometers long. | answer = 4.8 * 25000 / 100000 | 1.2 | rates | a map scale | a utility bill | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wce5ca8cd7aa1c3_3 | On Tuesday, Jolanta’s dog-walking shift runs from 8:35 AM to 2:20 PM, with an unpaid break of 28 minutes. How much of the shift is paid? | [
"8:35 AM to 2:20 PM = 5 hours 45 minutes",
"5 hours 45 minutes − 28 minutes = 5 hours 17 minutes",
"5 × 60 + 17 = 317 minutes"
] | 317 minutes | Jolanta’s paid shift is 317 minutes. | answer = (5 * 60 + 45) - 28 | 317 | clock | a shift with an unpaid break | a dog walking business | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wce5ca8cd7aa1c3_4 | On Friday, Patric’s bees travel at an average speed of 0.84 kilometers per hour while he tracks them across a field. He follows them for 2 hours and 35 minutes. How far do they travel? | [
"2 × 60 + 35 = 155 minutes",
"155 ÷ 60 = 31/12 hours",
"0.84 × 31/12 = 2.17 kilometers"
] | 2.17 kilometers | The bees travel 2.17 kilometers. | answer = round(0.84 * (2 + 35 / 60), 2) | 2.17 | rates | distance from speed and time | a beekeeper | batch-1 | 3 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wce5ca8cd7aa1c3_5 | $1,185 is Aslak and Susi’s current monthly rent, and their landlord is raising it by 4.6%. What will their new monthly rent be? | [
"$1,185 × 0.046 = $54.51",
"$1,185 + $54.51 = $1,239.51"
] | $1,239.51 | Their new monthly rent will be $1,239.51. | answer = round(1185 * (1 + 0.046), 2) | 1239.51 | percent | a percent increase | a rent payment | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wce5ca8cd7aa1c3_6 | If Petre’s hotel stay costs $684.00 before tax and the final bill is $739.62, how much of the pre-tax cost is the tax, expressed as a percent? | [
"$739.62 − $684.00 = $55.62 tax",
"$55.62 ÷ $684.00 = 0.0813…",
"0.0813… × 100 = 8.13…%"
] | 8.13% | The tax is about 8.13% of the pre-tax cost. | answer = round((739.62 - 684.00) / 684.00 * 100, 2) | 8.13 | percent | what percent one amount is of another | a hotel stay | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wce5ca8cd7aa1c3_7 | To calculate Martim’s weekly pay at the print shop, Niyazi notes that he earns $18.75 per hour and gets 1.5 times that rate for hours beyond 40. This week Martim worked 43.5 hours. How much did he earn before deductions? | [
"43.5 − 40 = 3.5 overtime hours",
"40 × $18.75 + 3.5 × ($18.75 × 1.5) = $848.44"
] | $848.44 | Martim earned $848.44 before deductions. | answer = round(40 * 18.75 + (43.5 - 40) * 18.75 * 1.5, 2) | 848.44 | money | weekly pay with overtime | a print shop | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w44d6e2765758b3_0 | If Dorottya's piano lesson begins on April 17 and that date is counted as day 1, Amaya wants to know when the 12th day falls. What is the date? | [
"April 17 is day 1",
"12 − 1 = 11 days after April 17",
"April 17 + 11 days = April 28"
] | April 28 | The 12th day falls on April 28. | from datetime import date, timedelta
answer = (date(2025, 4, 17) + timedelta(days=12 - 1)).strftime("%B %-d") | April 28 | dates | inclusive counting where the first day counts as day 1 | a piano lesson | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w44d6e2765758b3_1 | Gesine pays $30.00 for print-shop orders totaling $18.65. How much change should she receive? | [
"$30.00 − $18.65 = $11.35"
] | $11.35 | Gesine should receive $11.35 in change. | answer = round(30.00 - 18.65, 2) | 11.35 | money | change from a payment | a print shop | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w44d6e2765758b3_2 | On Wednesday, Keyvan and Aurel's quilting circle starts at 9:18 AM and ends at 12:07 PM. How long does the circle last? | [
"9:18 AM to 12:00 PM = 2 hours 42 minutes",
"12:00 PM to 12:07 PM = 7 minutes",
"162 + 7 = 169 minutes"
] | 169 minutes | The quilting circle lasts 169 minutes. | from datetime import datetime
start = datetime.strptime("9:18 AM", "%I:%M %p")
end = datetime.strptime("12:07 PM", "%I:%M %p")
answer = int((end - start).total_seconds() / 60) | 169 | clock | elapsed time between two clock times | a quilting circle | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w44d6e2765758b3_3 | To equip its robotics team, Ludwik and Kornelis can buy three bulk packs of eight sensors for $46.80 per pack, or buy all the sensors individually at $6.35 each. How much does the team save by choosing the bulk packs? | [
"3 × 8 = 24 sensors",
"24 × $6.35 = $152.40",
"3 × $46.80 = $140.40",
"$152.40 − $140.40 = $12.00"
] | $12.00 | The team saves $12.00 by choosing the bulk packs. | answer = round(24 * 6.35 - 3 * 46.80, 2) | 12.0 | money | a bulk deal versus singles | a robotics team | batch-1 | 4 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w44d6e2765758b3_4 | When Benigno and Sebastien reserve garden plots, the community garden charges a fixed $12.60 fee plus $4.25 for each of 17 plots. What is their total charge? | [
"10 × $4.25 = $42.50",
"7 × $4.25 = $29.75; $42.50 + $29.75 = $72.25",
"$72.25 + $12.60 = $84.85"
] | $84.85 | Their total charge is $84.85. | answer = round(12.60 + 17 * 4.25, 2) | 84.85 | money | a fixed fee plus a per-unit charge | a community garden | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w44d6e2765758b3_5 | Because Predrag's utility bill is $86.40 after a 28% discount, Behrouz wants to find its original price. What was the bill before the discount? | [
"100% − 28% = 72% of the original price",
"72% = 0.72",
"$86.40 ÷ 0.72 = $120.00"
] | $120.00 | The bill's original price was $120.00. | answer = round(86.40 / (1 - 0.28), 2) | 120.0 | percent | the original price from a sale price and discount | a utility bill | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w44d6e2765758b3_6 | Tilin travels 42 kilometers in 35 minutes on the way to a science fair, then covers another 68 kilometers in 55 minutes. What is Tilin's average speed for the whole trip in kilometers per hour? | [
"42 + 68 = 110 kilometers",
"35 + 55 = 90 minutes",
"90 ÷ 60 = 1.5 hours",
"110 ÷ 1.5 = 73.333… kilometers per hour"
] | 73.33 kilometers per hour | Tilin's average speed for the trip is 73.33 kilometers per hour. | answer = round((42 + 68) / ((35 + 55) / 60), 2) | 73.33 | averages | average speed over two legs of a trip | a science fair | batch-1 | 4 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w44d6e2765758b3_7 | On Monday, Gianpiero's book club has 38 members; by the next Monday, Takao reports that membership has grown to 47. What is the percentage increase, rounded to one decimal place? | [
"47 − 38 = 9 members",
"(9 ÷ 38) × 100 = 23.7%"
] | 23.7% | The book club's membership increased by 23.7%. | answer = round(((47 - 38) / 38) * 100, 1) | 23.7 | percent | a percent increase | a book club | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wc7164d34f990a4_0 | Two classroom supply orders are packed overnight by Folke. Packing begins at 10:48 PM and finishes at 1:23 AM the next day. How long does packing take? | [
"12:00 AM − 10:48 PM = 1 hour 12 minutes; 1 hour 12 minutes + 1 hour 23 minutes = 2 hours 35 minutes"
] | 2 hours 35 minutes | Packing takes 2 hours 35 minutes. | answer = 60 - (22 * 60 + 48 - 24 * 60) + (1 * 60 + 23) - 60 | 155 | clock | a time span that crosses midnight | a classroom supply order | batch-1 | 0 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wc7164d34f990a4_2 | For the quilting circle’s spring display, Evren and Edgard count 40 finished squares, and 35% of them have flower patterns. How many squares have flower patterns? | [
"35 ÷ 100 = 0.35",
"0.35 × 40 = 14 squares"
] | 14 squares | There are 14 flower-patterned squares. | answer = 40 * 35 / 100 | 14.0 | percent | a percent of a group | a quilting circle | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wc7164d34f990a4_3 | The laundromat shift for Petteri runs from 7:15 AM to 3:42 PM, with a 37-minute unpaid break. How much paid time is in the shift? | [
"3:42 PM − 7:15 AM = 8 hours 27 minutes",
"8 × 60 + 27 = 507 minutes",
"507 − 37 = 470 minutes = 7 hours 50 minutes"
] | 7 hours 50 minutes | Petteri has 7 hours 50 minutes of paid time. | answer = round(((15 * 60 + 42) - (7 * 60 + 15) - 37) / 60, 2) | 7.83 | clock | a shift with an unpaid break | a laundromat | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wc7164d34f990a4_4 | Every babysitting shift Sixten works is scheduled in Denver time, which is two hours behind Roderich’s time in New York. Sixten starts at 7:35 PM Denver time. What time is it in New York when the shift starts? | [
"7:35 PM + 2 hours = 9:35 PM",
"9:35 PM stays on the same day"
] | 9:35 PM | It is 9:35 PM in New York when the shift starts. | answer = '9:35 PM' | 9:35 PM | clock | a fixed time-zone offset | a babysitting job | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wc7164d34f990a4_5 | After the county fair ends on May 26, 2025, Emmi and Burchard look back to the date exactly five weeks earlier. What date was that? | [
"5 × 7 = 35 days",
"May 26 − 26 days = April 30",
"35 − 26 = 9 days",
"April 30 − 9 days = April 21"
] | April 21 | The date five weeks before the fair ended was April 21. | from datetime import date, timedelta
answer = (date(2025, 5, 26) - timedelta(weeks=5)).strftime('%B %-d') | April 21 | dates | a date N weeks earlier | a county fair | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wc7164d34f990a4_6 | At the end of a babysitting job, Josette and Johanne split a $38.17 payment evenly among four sitters. How much should each sitter receive if the leftover cent goes to one sitter? | [
"$38.17 = 3,817 cents",
"3,817 ÷ 4 = 954 cents each, remainder 1 cent",
"3 sitters receive $9.54 each; 1 sitter receives $9.55"
] | $9.54 each, with one sitter receiving $9.55 | Three sitters receive $9.54 each, and one receives $9.55. | answer = round((3817 // 4 + 1) / 100, 2) | 9.55 | money | splitting a cost evenly and handling the leftover cents | a babysitting job | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wc7164d34f990a4_7 | The medication schedule for Fikret and Tuomo lasts 17 days, counting the first day as day 1. If day 1 is March 4, 2025, what is the date of the final day? | [
"17 − 1 = 16 days after the start date",
"16 ÷ 7 = 2 weeks and 2 days",
"March 4 + 16 days = March 20"
] | March 20 | The final day of the medication schedule is March 20. | from datetime import date, timedelta
answer = (date(2025, 3, 4) + timedelta(days=17 - 1)).strftime('%B %-d') | March 20 | dates | inclusive counting where the first day counts as day 1 | a medication schedule | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w0f82dc983ad2cd_0 | The car wash fundraiser run by Ayako and Alen offers a wash priced at $18.75 with a 20% discount. The discounted price is then charged 8% sales tax. What is the final price? | [
"Discount: $18.75 × 0.20 = $3.75",
"Discounted price: $18.75 − $3.75 = $15.00",
"Tax-included price: $15.00 × 1.08 = $16.20"
] | $16.20 | The final price is $16.20. | answer = round(18.75 * (1 - 0.20) * (1 + 0.08), 2) | 16.2 | percent | a discount followed by sales tax | a car wash fundraiser | batch-1 | 3 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w0f82dc983ad2cd_1 | Since Yukihiro was born on May 21, 1998, the gym needs to know his age on August 14, 2026, to set up his membership. How old is he on that date? | [
"2026 − 1998 = 28 years"
] | 28 years | Yukihiro is 28 years old on August 14, 2026. | answer = 2026 - 1998 | 28 | dates | age on a given date from a birthday | a gym membership | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w0f82dc983ad2cd_2 | To plan their fishing trip, Daan and Heribert leave the dock at 6:35 AM and travel for 3 hours and 25 minutes. What time do they arrive? | [
"6:35 AM + 3 hours 25 minutes = 10:00 AM"
] | 10:00 AM | They arrive at 10:00 AM. | from datetime import datetime, timedelta
answer = (datetime.strptime('6:35 AM', '%I:%M %p') + timedelta(hours=3, minutes=25)).strftime('%I:%M %p').lstrip('0') | 10:00 AM | clock | add a duration to a start time | a fishing trip | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w0f82dc983ad2cd_3 | While Alin and Sibylle work in a small office, they prepare 24 information packets per hour for 7.25 hours. How many packets do they prepare? | [
"24 × 7.25 = 174 packets"
] | 174 packets | They prepare 174 packets. | answer = 24 * 7.25 | 174.0 | multistep | a production schedule with rates and hours | a small office | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w0f82dc983ad2cd_4 | Each lunch in the hospital cafeteria costs Edgard $8.35 for a sandwich and $4.60 for soup, and he also buys two coffees at $2.15 each. He pays with a $25 bill. How much change does he get? | [
"Coffee cost: 2 × $2.15 = $4.30",
"Sandwich and soup: $8.35 + $4.60 = $12.95",
"Total cost: $12.95 + $4.30 = $17.25",
"Change: $25.00 − $17.25 = $7.75"
] | $7.75 | Edgard gets $7.75 in change. | answer = round(25.00 - (8.35 + 4.60 + 2 * 2.15), 2) | 7.75 | money | change from a payment | a hospital cafeteria | batch-1 | 4 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w0f82dc983ad2cd_6 | October 9, 2026, is the date Toshio and Gerda begin a newspaper route. They want to check the route notes from exactly five weeks earlier. What date are they looking for? | [
"October 9, 2026 − 5 weeks = September 4, 2026"
] | September 4 | They are looking for the notes dated September 4. | from datetime import date, timedelta
answer = (date(2026, 10, 9) - timedelta(weeks=5)).strftime('%B %-d') | September 4 | dates | a date N weeks earlier | a newspaper route | batch-1 | 0 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w0f82dc983ad2cd_7 | Three parts of the family reunion schedule last 2 hours 35 minutes, 1 hour 48 minutes, and 52 minutes. Kambiz and Cathrin want to know the total duration. How long is the schedule altogether? | [
"2 × 60 + 35 + 1 × 60 + 48 + 52 = 315 minutes",
"315 minutes = 5 hours 15 minutes"
] | 5 hours 15 minutes | The schedule lasts 5 hours 15 minutes altogether. | answer = (2 * 60 + 35 + 1 * 60 + 48 + 52) / 60 | 5.25 | durations | add several durations in hours and minutes | a family reunion | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wf5b7fc820dd662_0 | At the community garden, Georg and Klavdiya harvested 2.35 kg of tomatoes and 780 g of peppers. How many kilograms of produce did they harvest altogether? | [
"2.35 × 1,000 = 2,350 g",
"2,350 + 780 = 3,130 g",
"3,130 ÷ 1,000 = 3.13 kg"
] | 3.13 kg | They harvested 3.13 kg of produce altogether. | answer = round((2.35 * 1000 + 780) / 1000, 2) | 3.13 | units | metric length or mass conversion | a community garden | batch-1 | 3 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wf5b7fc820dd662_1 | $425 was raised at the recycling drive before Otfried and Severino increased the total by 12% with a final collection push. How much money had been raised after the increase? | [
"12% = 0.12",
"$425 × 0.12 = $51",
"$425 + $51 = $476"
] | $476.00 | The recycling drive had raised $476.00 after the increase. | answer = round(425 + 425 * 0.12, 2) | 476.0 | percent | a percent increase | a recycling drive | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wf5b7fc820dd662_3 | Each day at summer camp, Brunon and Neill use 27 snack packs. The camp starts with 213 packs and runs for 5 days. How many packs remain after those days? | [
"27 × 5 = 135 packs used",
"213 − 135 = 78 packs"
] | 78 packs | There are 78 snack packs remaining. | answer = 213 - 27 * 5 | 78 | multistep | inventory changing over several days | a summer camp | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wf5b7fc820dd662_4 | If Urszula arrives at the dentist office at 9:18 AM and leaves at 12:07 PM, how long is her visit? | [
"9:18 AM to 10:00 AM = 42 minutes",
"10:00 AM to 12:00 PM = 120 minutes",
"12:00 PM to 12:07 PM = 7 minutes",
"42 + 120 + 7 = 169 minutes"
] | 169 minutes | Urszula's visit lasts 169 minutes. | from datetime import date, datetime, timedelta
answer = (datetime.strptime("12:07 PM", "%I:%M %p") - datetime.strptime("9:18 AM", "%I:%M %p")).seconds // 60 | 169 | clock | elapsed time between two clock times | a dentist office | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wf5b7fc820dd662_5 | Three summer camp groups, led by Otmar and Yuval, sold $684, $539, and $777 worth of crafts. The camp pays a 6% commission on the combined sales. How much commission will it pay? | [
"$684 + $539 + $777 = $2,000",
"6% = 0.06",
"$2,000 × 0.06 = $120",
"Commission = $120.00"
] | $120.00 | The camp will pay $120.00 in commission. | answer = round((684 + 539 + 777) * 0.06, 2) | 120.0 | percent | a commission on sales | a summer camp | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
wf5b7fc820dd662_7 | During a phone-plan promotion, Carolien and Halyna set a reminder for 9 days after the plan begins on August 26. What date should they set the reminder for? | [
"August 26 + 5 days = August 31",
"9 − 5 = 4 days remaining",
"August 31 + 4 days = September 4"
] | September 4 | They should set the reminder for September 4. | from datetime import date, datetime, timedelta
answer = (date(2025, 8, 26) + timedelta(days=9)).strftime("%B %-d") | September 4 | dates | N days after a date across a month end | a phone plan | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w88afc414563277_1 | A county fair opens its main stage at 2:18 PM, and the final show ends at 6:07 PM. Tanja and Silvie stay for the entire time. How long are they at the stage? | [
"2:18 PM to 3:00 PM = 42 minutes",
"3:00 PM to 6:00 PM = 180 minutes",
"6:00 PM to 6:07 PM = 7 minutes; 42 + 180 + 7 = 229 minutes = 3 hours 49 minutes"
] | 3 hours 49 minutes | Tanja and Silvie are at the stage for 3 hours 49 minutes. | answer = (3 * 60 + 49) | 229 | clock | elapsed time between two clock times | a county fair | batch-1 | 0 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w88afc414563277_2 | A used bookstore has a novel priced at $17.65 and a bookmark priced at $2.35. Zuzana pays with a $25 bill, then spends 23 minutes browsing before checking out. If she arrives at 1:42 PM, when does she leave? | [
"$17.65 + $2.35 = $20.00",
"$25.00 − $20.00 = $5.00 change",
"1:42 PM + 23 minutes = 2:05 PM"
] | 2:05 PM | Zuzana leaves the bookstore at 2:05 PM. | from datetime import datetime, timedelta
answer = (datetime.strptime('1:42 PM', '%I:%M %p') + timedelta(minutes=23)).strftime('%I:%M %p').lstrip('0') | 2:05 PM | multistep | three steps mixing money and time | a used bookstore | batch-1 | 3 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w88afc414563277_3 | To compare two bakery deals, Erling and Oleksandr check a box of 8 rolls for $13.52 and a box of 6 rolls for $10.50. Which box has the lower price per roll, and what is that price? | [
"$13.52 ÷ 8 = $1.69 per roll",
"$10.50 ÷ 6 = $1.75 per roll",
"$1.75 − $1.69 = $0.06; the 8-roll box is cheaper by $0.06 per roll"
] | 8-roll box, $1.69 per roll | The 8-roll box has the lower price, at $1.69 per roll. | answer = round(min(13.52 / 8, 10.50 / 6), 2) | 1.69 | money | comparing unit prices | a bakery | batch-1 | 3 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w88afc414563277_5 | On Saturday at the county fair, Dorte sees the fireworks listed for 20:35 on the schedule. Pille wants to write the time in the 12-hour format. What time should she write? | [
"20 − 12 = 8",
"The minutes remain 35",
"20:35 = 8:35 PM"
] | 8:35 PM | Pille should write 8:35 PM. | hour = 20 - 12
answer = f'{hour}:{35:02d} PM' | 8:35 PM | clock | a 24-hour time converted to a 12-hour time inside the problem | a county fair | batch-1 | 1 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w88afc414563277_6 | At the dentist office, Yoshie records four appointment sessions for Furio during one week: 38 minutes, 42 minutes, 35 minutes, and 47 minutes. How much time do the sessions take altogether? | [
"38 + 42 = 80 minutes",
"35 + 47 = 82 minutes",
"80 + 82 = 162 minutes",
"162 minutes = 2 hours 42 minutes"
] | 2 hours 42 minutes | Furio’s sessions take 2 hours 42 minutes altogether. | answer = 38 + 42 + 35 + 47 | 162 | durations | total time over several sessions in a week | a dentist office | batch-1 | 3 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w88afc414563277_7 | On Sunday at a holiday market, Ahti finds a scarf originally priced at $46.80. Latifa tells him it is marked down by 18%. What is the new price? | [
"$46.80 × 0.18 = $8.424, which rounds to $8.42 off",
"$46.80 − $8.42 = $38.38"
] | $38.38 | The scarf’s new price is $38.38. | answer = round(46.80 * (1 - 0.18), 2) | 38.38 | percent | a percent decrease | a holiday market | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w8060a80577710d_0 | For Kati's birthday, a power outage forces her and Masakatsu to rent a backup generator for $38.40. They have 14 guests, and food costs $18.75 per person while glow sticks cost $4.60 per person. How much will the party cost altogether? | [
"$18.75 + $4.60 = $23.35 per person",
"14 × $23.35 = $326.90",
"$326.90 + $38.40 = $365.30"
] | $365.30 | The party will cost $365.30 altogether. | answer = round(14 * (18.75 + 4.60) + 38.40, 2) | 365.3 | multistep | a party budget with per-person costs | a power outage | batch-1 | 3 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
w8060a80577710d_1 | Since Gottlieb and Grethe's hiking club is buying a $86.50 trail map set, the shop offers 20% off and then an additional 12% off the reduced price. What is the final price? | [
"$86.50 × 0.80 = $69.20",
"$69.20 × 0.88 = $60.896 ≈ $60.90"
] | $60.90 | The hiking club will pay $60.90. | answer = round(86.50 * 0.80 * 0.88, 2) | 60.9 | percent | two successive discounts | a hiking club | batch-1 | 2 | gpt-6-luna (flex, low reasoning effort) | sandboxed execution (no network, CPU/memory/time limits); output equals stated answer and final sentence; arithmetic/clock steps re-computed | CC-BY-4.0 |
End of preview. Expand in Data Studio
OpenLemma v0.4: word problems with executed programs
This dataset holds everyday word problems. Each one comes with short worked steps and a Python program. Every program was executed in a sandbox, and its output matches the stated answer and the final sentence. It is the word-problem family of OpenLemma v0.4's natural-language-to-program items, and it is used in ALEA model training.
45,112 problems. manifest.json gives the breakdown by type and origin, and file checksums.
Columns
| column | content |
|---|---|
id |
row id |
problem |
the problem text |
steps |
the worked steps; every arithmetic and clock-time step was re-computed |
answer, final |
the final answer, and the sentence that states it |
program |
Python 3; plain arithmetic and datetime only |
executed_output |
what the program printed when run |
type |
one of 15 types: clock, dates, durations, percent, money, products, quotients, differences, fractions, geometry, rates, tiers, averages, units, multistep |
subtype, setting |
the problem's subtype and its setting (about 270 everyday and small-business settings) |
origin |
generation batch |
steps_recomputed |
number of steps that were re-computed |
generator |
the model that wrote the problem |
verification |
how the row was verified |
license |
row license |
Generation
- Generator: gpt-6-luna (OpenAI, flex tier, low reasoning effort).
- Prompts: fully synthetic. Each one is a list of specs: problem type, setting, opening style, number of steps and two given names.
- Excluded settings: no court, legal, medical or police settings.
- Safety screen: every prompt and output passed a fail-closed screen for violence, crime, sexual content, weapons, drugs and similar topics.
Verification
All checks fail closed.
- Static check: the program is parsed and screened. No input/output, no imports other than
datetime, no while loops, no lambdas, no dunder access. A program that usesdateordatetimewithout importing it gets that one import line added. - Step re-computation: every machine-checkable step (arithmetic, clock times) is re-computed, and a single false step rejects the problem.
- Sandboxed execution:
- a fresh isolated interpreter with no network access;
- limits of 4 s CPU, 1 GiB memory and 8 s wall time;
- an empty environment.
- Answer match: the executed answer must equal the stated answer and the final sentence, within the precision shown. Answers that give fractional counts of discrete things are rejected. So are weekday answers when the problem gives no year.
Decontamination
Problems are dropped if they overlap any of the following:
- GSM8K (train and test word 8-grams and number signatures) or ALEA's reasoning prompts;
- the ALEA evaluation suites;
- the arithmetic-notation probe, including its held-out operand pairs;
- a set of copyrighted-text probes.
Release screens
- Every field of every row was checked again against ALEA's held-out probes (held-out operand and word pairs); the 751 rows with a hit were removed.
- Toxicity (
s-nlp/roberta_toxicity_classifier, every 64-word chunk): no row at 0.8 or above. - No private contact details (e-mail, phone, street address) were found.
License
CC-BY-4.0.
No evaluation data is included: ALEA's held-out evaluation sets are kept private. The generation and verification code will be released with the ALEA model.
- Downloads last month
- 37
Models trained or fine-tuned on mjbommar/openlemma-v0.4-wp