Faceted search with filters & sorting
Use site filters, sorts and comparisons to find the item meeting all constraints.
20 project-authored proposals · taxonomy v1.1 · 0 new benchmark runs. Verifiers and partial-credit rules below are designs, not executed results.
- Typical applications
- shopping and listings sites
- Class GUI-need label
G2· GUI-essential
Task proposals · 20
Expand to inspect the complete original wording and proposed checks. Original proposal text remains English in both interfaces.
R04.01.02.01-T01 Cheapest 4K USB-C monitor meeting specs
Original instruction
On the replica store 'Omnizon' (fixed snapshot), use the filters to find the cheapest 27-inch 4K monitor with USB-C power delivery >= 65 W and an average rating >= 4.3 that can be delivered by the date shown in task.txt. Write the product ID to answer.txt.
- Platforms
web· Any OS (web)- Applications
- shopping and listings sites
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clickvis· Visual judgment- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
answer.txt product ID equals the unique cheapest item satisfying all filters on the snapshot. Must be reached via the site's filter/sort UI on the fixed catalog; the backend is deterministic. Omnizon issues no API token and serves catalog data only to browser sessions (session cookie + CSRF); its request log must show the filter/sort requests that reach the answer.
Proposed partial credit
40% filters applied correctly; 30% sorted by price; 30% exact product ID
R04.01.02.01-T02 Newest pet-friendly 2-bed rentals
Original instruction
On the replica listings site 'Zilloft', filter for 2-bedroom rentals under $2,800/month that allow cats in ZIP 94110, sort by newest first, and write the first three listing IDs in order to answer.txt.
- Platforms
web· Any OS (web)- Applications
- shopping and listings sites
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clickvis· Visual judgment- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
answer.txt lists the three correct listing IDs in the exact newest-first order from the snapshot. Reached via the site filters/sort; deterministic backend. Zilloft issues no API token and serves catalog data only to browser sessions (session cookie + CSRF); its request log must show the filter/sort requests that reach the answer.
Proposed partial credit
40% correct filtered set; 30% correct sort; 30% exact ordered top-3
R04.01.02.01-T03 Short-stay booking within budget
Original instruction
On the replica 'Staynb', find a stay for 4 guests, 3 nights from the date in task.txt, that has a washer and self check-in, with a total (including all fees) <= $900. Write the listing ID and the total to answer.txt.
- Platforms
web· Any OS (web)- Applications
- shopping and listings sites
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clickvis· Visual judgment- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
answer.txt listing ID is a valid option meeting all constraints and the total matches the snapshot's computed total (incl. fees) within $1. Reached via the UI; deterministic. Staynb issues no API token and serves catalog data only to browser sessions (session cookie + CSRF); its request log must show the search and filter requests that reach the answer.
Proposed partial credit
40% amenity+occupancy filters; 30% total within budget; 30% exact listing+total
R04.01.02.01-T04 Top earbuds by review count
Original instruction
On 'Omnizon', filter for noise-cancelling earbuds with battery life >= 30 hours released in 2025 or later, sort by number of reviews (descending), and write the top item's SKU to answer.txt.
- Platforms
web· Any OS (web)- Applications
- shopping and listings sites
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clickvis· Visual judgment- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
answer.txt SKU equals the most-reviewed item meeting the filters on the snapshot. Reached via UI filter/sort; deterministic. Omnizon issues no API token and serves catalog data only to browser sessions (session cookie + CSRF); its request log must show the filter/sort requests that reach the answer.
Proposed partial credit
40% filters; 30% sort by reviews; 30% exact SKU
R04.01.02.01-T05 In-stock laptop under budget
Original instruction
On 'Omnizon', find the laptop with at least 32 GB RAM priced under $1,200 with rating >= 4.5 that is currently in stock. If several qualify, pick the highest-rated, then cheapest. Write the product ID to answer.txt.
- Platforms
web· Any OS (web)- Applications
- shopping and listings sites
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clickvis· Visual judgment- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
answer.txt product ID equals the item selected by the stated tie-break on the snapshot. Reached via UI filter/sort; deterministic. Omnizon issues no API token and serves catalog data only to browser sessions (session cookie + CSRF); its request log must show the filter/sort requests that reach the answer.
Proposed partial credit
40% filters incl. stock; 30% tie-break applied; 30% exact product ID
R04.01.02.01-T06 Cheapest nonstop within a time window
Original instruction
On the replica travel site 'Flylott', search the route and dates in task.txt, filter to nonstop flights under 6 hours departing in the morning window, sort by price, and write the cheapest matching flight's ID to answer.txt. Do not book.
- Platforms
web· Any OS (web)- Applications
- shopping and listings sites
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clickvis· Visual judgment- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
answer.txt flight ID equals the cheapest nonstop under 6h in the window on the snapshot. Reached via UI filters/sort; deterministic; no booking performed. Flylott issues no API token and serves catalog data only to browser sessions (session cookie + CSRF); its request log must show the filter/sort requests that reach the answer.
Proposed partial credit
40% nonstop+duration+window filters; 30% price sort; 30% exact flight ID
R04.01.02.01-T07 Used car within distance and budget
Original instruction
On the replica marketplace 'Autolott', filter used cars with under 60,000 miles, automatic transmission, priced under $20,000, within 50 miles of the given ZIP, sorted by lowest price. Write the top listing's ID to answer.txt.
- Platforms
web· Any OS (web)- Applications
- shopping and listings sites
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clickvis· Visual judgment- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
answer.txt listing ID equals the cheapest car meeting all filters on the snapshot. Reached via UI filters/sort incl. distance; deterministic. Autolott issues no API token and serves catalog data only to browser sessions (session cookie + CSRF); its request log must show the filter (incl. distance) requests that reach the answer.
Proposed partial credit
40% all filters incl. distance; 30% price sort; 30% exact listing ID
R04.01.02.01-T08 Side-by-side compare to pick a camera
Original instruction
On 'Omnizon', shortlist cameras under $900 with 4K60 video, in-body stabilization, and a tilting screen, add up to three to the Compare view, and identify the one that also has the longest battery life. Write its product ID to answer.txt.
- Platforms
web· Any OS (web)- Applications
- shopping and listings sites
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clickvis· Visual judgmenthover· Hover & context- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
answer.txt product ID equals the shortlisted camera with the max battery life meeting all constraints on the snapshot; the Compare view was used. Deterministic. Omnizon issues no API token and serves catalog data only to browser sessions (session cookie + CSRF); its request log must show the filter and Compare-view requests that reach the answer.
Proposed partial credit
40% correct shortlist; 30% compare view used; 30% exact product ID
R04.01.02.01-T09 Hotel with cancellation and breakfast
Original instruction
On 'Staynb-Hotels', find hotels for the dates in task.txt with free cancellation, breakfast included, guest rating >= 8.5, under $200/night, sorted by price. Write the cheapest qualifying hotel's ID and its nightly price to answer.txt.
- Platforms
web· Any OS (web)- Applications
- shopping and listings sites
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clickvis· Visual judgment- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
answer.txt hotel ID and nightly price match the cheapest qualifying hotel on the snapshot. Reached via UI filters/sort; deterministic. Staynb-Hotels issues no API token and serves catalog data only to browser sessions (session cookie + CSRF); its request log must show the filter/sort requests that reach the answer.
Proposed partial credit
40% filters; 30% price sort; 30% exact hotel+price
R04.01.02.01-T10 Best per-unit grocery price
Original instruction
On the replica grocery 'Freshlott', among the three brands listed in task.txt for a product, filter to the given size range and determine which specific SKU has the lowest price per unit (per 100g). Write that SKU to answer.txt.
- Platforms
web· Any OS (web)- Applications
- shopping and listings sites
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clickvis· Visual judgment- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
answer.txt SKU equals the lowest per-unit option among the filtered brands on the snapshot (per-unit computed from price/size). Deterministic. Freshlott issues no API token and serves catalog data only to browser sessions (session cookie + CSRF); its request log must show the brand/size filter requests that reach the answer.
Proposed partial credit
40% brand+size filters; 30% per-unit comparison; 30% exact SKU
R04.01.02.01-T11 Recent remote jobs above a salary
Original instruction
On the replica board 'Jobbly', filter for remote postings with salary >= the amount in task.txt posted in the last 7 days in a named category, sort by most recent, and write the count and the first three job IDs to answer.json.
- Platforms
web· Any OS (web)- Applications
- shopping and listings sites
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clickvis· Visual judgment- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
answer.json count equals the number of matching postings and the three IDs are the most-recent ones in order on the snapshot. Reached via UI filters/sort; deterministic. Jobbly issues no API token and serves catalog data only to browser sessions (session cookie + CSRF); its request log must show the filter/sort requests that reach the answer.
Proposed partial credit
40% filters; 30% correct count; 30% top-3 ordered IDs
R04.01.02.01-T12 Simple multi-filter shoe search
Original instruction
On 'Omnizon', filter men's shoes in size 10, waterproof, color black, priced under $150, sorted by rating. Write the top result's product ID to answer.txt.
- Platforms
web· Any OS (web)- Applications
- shopping and listings sites
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clickvis· Visual judgment- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 2 / 5
Proposed verifier
answer.txt product ID equals the highest-rated shoe meeting all filters on the snapshot. Deterministic. Omnizon issues no API token and serves catalog data only to browser sessions (session cookie + CSRF); its request log must show the filter/sort requests that reach the answer.
Proposed partial credit
50% filters applied; 20% rating sort; 30% exact product ID
R04.01.02.01-T13 Amenity + map-bounded apartment search
Original instruction
On 'Zilloft', find a 1-bed apartment with in-unit laundry, parking, and a gym, located inside the map boundary shown in task.txt (use the map's draw/boundary filter), under $2,500. Write the listing ID to answer.txt.
- Platforms
web· Any OS (web)- Applications
- shopping and listings sites
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clickvis· Visual judgmentdrag· Drag & drop- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
answer.txt listing ID is within the specified map boundary and meets all amenity/price filters on the snapshot. Reached via the map-area filter + UI; deterministic. Zilloft issues no API token and serves catalog data only to browser sessions (session cookie + CSRF); its request log must show the map-boundary filter requests that reach the answer.
Proposed partial credit
30% amenity filters; 30% map-boundary applied; 40% exact listing ID
R04.01.02.01-T14 Warranty filter with brand exclusion
Original instruction
On 'Omnizon', find power tools with a 2+ year warranty, excluding a named brand, sorted by price ascending, and write the cheapest qualifying product ID to answer.txt.
- Platforms
web· Any OS (web)- Applications
- shopping and listings sites
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clickvis· Visual judgment- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
answer.txt product ID equals the cheapest item with 2+ year warranty excluding the named brand on the snapshot. Deterministic. Omnizon issues no API token and serves catalog data only to browser sessions (session cookie + CSRF); its request log must show the filter/sort requests that reach the answer.
Proposed partial credit
40% warranty+exclusion filters; 30% price sort; 30% exact product ID
R04.01.02.01-T15 Specific book edition in stock
Original instruction
On the replica 'Booklott', find the hardcover edition of the title in task.txt priced under $30 that is in stock. Write its product ID to answer.txt.
- Platforms
web· Any OS (web)- Applications
- shopping and listings sites
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clickvis· Visual judgment- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 2 / 5
Proposed verifier
answer.txt product ID equals the correct hardcover edition under $30 in stock on the snapshot. Deterministic. Booklott issues no API token and serves catalog data only to browser sessions (session cookie + CSRF); its request log must show the search and filter requests that reach the answer.
Proposed partial credit
50% format+price+stock filters; 50% exact edition ID
R04.01.02.01-T16 Furniture within dimension limits
Original instruction
On 'Omnizon-Home', find a bookshelf no wider than 80 cm and no taller than 180 cm, in white, that does not require assembly, sorted by price. Write the top matching product ID to answer.txt.
- Platforms
web· Any OS (web)- Applications
- shopping and listings sites
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clickvis· Visual judgment- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
answer.txt product ID equals the cheapest bookshelf within the dimension/color/assembly constraints on the snapshot. Deterministic. Omnizon-Home issues no API token and serves catalog data only to browser sessions (session cookie + CSRF); its request log must show the filter/sort requests that reach the answer.
Proposed partial credit
40% dimension+color+assembly filters; 30% price sort; 30% exact product ID
R04.01.02.01-T17 Top-rated beginner short course
Original instruction
On the replica catalog 'Learnlott', filter courses rated >= 4.7, under 10 hours, at beginner level in a named subject, sorted by rating. Write the top course ID to answer.txt.
- Platforms
web· Any OS (web)- Applications
- shopping and listings sites
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clickvis· Visual judgment- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 2 / 5
Proposed verifier
answer.txt course ID equals the highest-rated course meeting the filters on the snapshot. Deterministic. Learnlott issues no API token and serves catalog data only to browser sessions (session cookie + CSRF); its request log must show the filter/sort requests that reach the answer.
Proposed partial credit
50% filters; 20% rating sort; 30% exact course ID
R04.01.02.01-T18 Single item meeting five filters
Original instruction
On 'Omnizon', exactly one refrigerator matches all five constraints in task.txt (counter-depth, French door, energy rating A, width <= 90 cm, under $2,000). Apply the filters to isolate it and write its product ID to answer.txt.
- Platforms
web· Any OS (web)- Applications
- shopping and listings sites
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clickvis· Visual judgment- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
answer.txt product ID equals the unique item matching all five filters on the snapshot (result set size 1). Deterministic. Omnizon issues no API token and serves catalog data only to browser sessions (session cookie + CSRF); its request log must show the filter/sort requests that reach the answer.
Proposed partial credit
50% all five filters applied; 20% result narrowed to one; 30% exact product ID
R04.01.02.01-T19 Best total including fees after sort
Original instruction
On 'Staynb', for the dates and occupancy in task.txt, the lowest nightly price is not the lowest total once cleaning/service fees are added. Sort by price, then compare the totals of the top five to find the genuinely cheapest total. Write that listing ID and total to answer.json.
- Platforms
web· Any OS (web)- Applications
- shopping and listings sites
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clickvis· Visual judgment- Work product
WP12· Answer/decision- Conditions
- None declared
- Proposed difficulty
- 5 / 5
Proposed verifier
answer.json listing ID is the one with the lowest all-in total among the qualifying set and the total matches the snapshot within $1. Deterministic; requires reading per-listing fee breakdowns in the UI. Staynb issues no API token and serves catalog data only to browser sessions (session cookie + CSRF); its request log must show the sort and fee-breakdown requests that reach the answer.
Proposed partial credit
30% qualifying set; 30% totals compared incl. fees; 40% exact cheapest listing+total
R04.01.02.01-T20 Best value across two result pages, then cart
Original instruction
On 'Omnizon', compare TVs across two filtered result pages (65-inch, 4K, rating >= 4.4) to find the best price-per-inch value, add it to the cart, and stop at the cart page without checking out. Write the chosen product ID to answer.txt.
- Platforms
web· Any OS (web)- Applications
- shopping and listings sites
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clickvis· Visual judgment- Work product
WP08· App/system state- Conditions
risk· Irreversible/risky- Proposed difficulty
- 5 / 5
Proposed verifier
answer.txt product ID equals the best price-per-inch TV meeting the filters across both pages on the snapshot, and the cart contains exactly that item with no order placed. Deterministic; stops at the cart boundary. Omnizon issues no API token and serves catalog data only to browser sessions (session cookie + CSRF); its request log must show the filter, paging and add-to-cart requests that reach the answer.
Proposed partial credit
30% filtered across both pages; 30% best value identified; 30% correct item in cart; 10% stopped before checkout
Mapped research references
Existing benchmark-to-class links from the research catalog. They are not record-level annotations, executions or product endorsements.