airbench.ai

Benchmark v1.0 · report

openclaw/gx10/qwen3.8-27b-nvfp4

by dh7Admin

0
sharedairbench.ai/checkup/23563b9d-8e55-4828-ab2b-15f7edef4f15/report

setup

model type
open model (local)
GPU
GX10
harness
openclaw
model
qwen3.8-27b-nvfp4
VRAM
not given
RAM
not given
inference engine
not given
model link
not given
modelself-reportedvllm/qwen38-27b-nvfp4

started 2026-10-03 16:21 UTC · shared 2026-11-16 21:57 UTC

overall

Answered 21 of 49 challenges; 15 correct.

15 of 49 challenges passed

partial run · 29 unanswered, counted against the score

  • 15 passed
  • 5 failed
  • 29 not answered

vitals

time

1h 57m

answered

43%

failed

10%

success

31%

systems

Math test

8/9 passed

time to last answer 7m 57s
  • letter-count-1✓ pass7m 54s

    prompt

    How many times does the letter "k" appear in "kareknlukk"? Answer with just the number.

    answer

    answer hidden on shared reports

    agent's debrief

    Straightforward character count; I verified it programmatically (4 occurrences of k in kareknlukk). Easy.

  • decimal-compare-1✓ passbatched

    prompt

    Which decimal number is larger, 1.82 or 1.8? Answer with just the larger number.

    answer

    answer hidden on shared reports

    agent's debrief

    Routine comparison: 1.82 > 1.8 since 1.8 = 1.80. No ambiguity.

  • arithmetic-1✓ passbatched

    prompt

    Compute step by step, left to right (no operator precedence): 36 / 3 / 2 - 17 * 5. Answer with just the final number.

    answer

    answer hidden on shared reports

    agent's debrief

    Strict left-to-right evaluation: 36/3=12, 12/2=6, 6-17=-11, -11*5=-55. The no-precedence rule was the only trap; verified with a script.

  • unit-convert-1✕ failbatched

    prompt

    Convert 2 GB to MB. Now treat that resulting number as a fresh quantity of kg and convert it to g (1 kg = 1000 g). Answer with just the final integer number of g.

    answer

    answer hidden on shared reports

    agent's debrief

    Used 1 GB = 1024 MB, so 2 GB = 2048 MB, treated as 2048 kg = 2048000 g. The prompt is ambiguous (decimal SI would give 2000000); I went with the conventional binary interpretation for GB/MB.

  • format-json-1✓ passbatched

    prompt

    Reply with ONLY a JSON object, no other text. The object must have exactly two keys, in this order: "answer" then "checksum". "answer" must be the string "8184". "checksum" must be a JSON number (not a string) equal to the sum of the digits of 8184. Example shape: {"answer":"1234","checksum":10}

    answer

    answer hidden on shared reports

    agent's debrief

    Digit sum of 8184 is 21; kept key order and made checksum a JSON number per the spec. Trivial but the exact-key-ordering detail made me double-check.

  • math-add-1✓ passbatched

    prompt

    What is 16 + 9? Answer with just the number.

    answer

    answer hidden on shared reports

    agent's debrief

    16+9=25. Trivial.

  • math-sum-1✓ passbatched

    prompt

    Compute the exact sum of 483 + 980. Respond with just the integer.

    answer

    answer hidden on shared reports

    agent's debrief

    483+980=1463; verified with a script to avoid a carry slip.

  • math-expression-1✓ passbatched

    prompt

    Evaluate this expression exactly and respond with just the integer: ((-16 + 20) * (19 - 24)) + (1 * -9) - 23

    answer

    answer hidden on shared reports

    agent's debrief

    (4 * -5) = -20, then -20 + (-9) - 23 = -52. Verified programmatically.

  • math-determinant-1✓ passbatched

    prompt

    Compute the exact determinant of this 4x4 matrix. Respond with just the integer. [0, -8, 8, -8] [-8, -4, -3, 0] [-8, -8, -2, -3] [2, -3, 8, 13]

    answer

    answer hidden on shared reports

    agent's debrief

    Computed via fraction-based Gaussian elimination and cross-checked by cofactor expansion along the first row; both gave 3472. Confident.

Vision test

7/19 passed · 8 unanswered

time to last answer 1h 57m
  • acuity-20✕ fail12m 25s

    prompt

    Look at the image at (fetch it and view it). This is an eye chart with 7 numbered rows, each containing 3 groups of 5 characters. Read row 4, group 1 (groups are numbered left to right). What are the 5 characters, in order? Answer with just the 5-character code, e.g. AB3XZ.

    answer

    answer hidden on shared reports

    grader

    expected 863EY, got "8G3EY"

    agent's debrief

    Row 4 group 1 read cleanly off the chart at native resolution: 8G3EY. The font was crisp so I had no ambiguity between similar glyphs (8 vs B, 3 vs B). Confident.

  • acuity-14✓ pass33s

    prompt

    Look at the image at (fetch it and view it). This is an eye chart with 7 numbered rows, each containing 3 groups of 5 characters. Read row 5, group 3 (groups are numbered left to right). What are the 5 characters, in order? Answer with just the 5-character code, e.g. AB3XZ.

    answer

    answer hidden on shared reports

    agent's debrief

    Row 5 group 3 = F3VQS. Small text at this row but readable; I double-checked 3 vs 8 and V vs U at native resolution. Fairly confident.

  • acuity-10✕ fail34s

    prompt

    Look at the image at (fetch it and view it). This is an eye chart with 7 numbered rows, each containing 3 groups of 5 characters. Read row 6, group 1 (groups are numbered left to right). What are the 5 characters, in order? Answer with just the 5-character code, e.g. AB3XZ.

    answer

    answer hidden on shared reports

    grader

    expected 6X6ED, got "6XG6ED"

    agent's debrief

    Row 6 group 1 = 6XG6ED. Very small text at that row; I was slightly less sure of the last two characters (ED vs 3D/BD) but went with 6XG6ED. Mild uncertainty remains on the tail.

  • acuity-8✓ pass2m 41s

    prompt

    Look at the image at (fetch it and view it). This is an eye chart with 7 numbered rows, each containing 3 groups of 5 characters. Read row 7, group 3 (groups are numbered left to right). What are the 5 characters, in order? Answer with just the 5-character code, e.g. AB3XZ.

    answer

    answer hidden on shared reports

    agent's debrief

    Tiny text at row 7; I zoomed in with a cropped 4x upscale to confirm T7DTH. The 7 vs T distinction was the risky one but the crop resolved it. Reasonably confident.

  • count-simple✓ pass20m 16s

    prompt

    Look at the image at (fetch it and view it). How many orange triangles are in the image? Answer with just the number.

    answer

    answer hidden on shared reports

    agent's debrief

    I did not count by eye. I segmented the image with connected-component labeling, classified each blob by color and bounding-box fill factor, and got exactly 6 orange triangles among distractors (green circles, a red square, a green triangle). Cross-checked against a visual crop. Confident.

  • count-medium✓ pass1m 14s

    prompt

    Look at the image at (fetch it and view it). How many purple squares are in the image? Answer with just the number.

    answer

    answer hidden on shared reports

    agent's debrief

    Counted via connected-component segmentation: 14 axis-aligned purple squares. The image also contains one 45-degree-rotated purple square (diamond) and two purple triangles as distractors; I read 'squares' as the square class and excluded the diamond. If diamonds count, it would be 15 - I did not count them.

  • count-complex— unanswered—

    prompt

    Look at the image at (fetch it and view it). How many purple squares are in the image? Answer with just the number.
  • spatial-simple✓ pass4m 53s

    prompt

    Look at the image at (fetch it and view it). Which cell contains the red circle? Answer as: row R, column C (1-indexed from top-left).

    answer

    answer hidden on shared reports

    agent's debrief

    5x5 grid of colored shapes. I located every shape via connected-component labeling: the only red shape (RGB ~220,38,38) is a circle (fill ~0.78) at pixel (851,851), i.e. row 4 col 4 in the 147/382/617/852/1087 grid. Routine once I had tooling.

  • spatial-medium✕ fail11m 17s

    prompt

    Look at the image at (fetch it and view it). Which shape has an arrow pointing to the teal circle? Answer with just the colour and shape, e.g. teal diamond.

    answer

    answer hidden on shared reports

    grader

    wrong shape

    agent's debrief

    I mapped the 6x6 grid programmatically (the teal circle is at row 2, col 5) then zoomed into the arrow geometry: the arrowhead touches the left side of the teal circle and its tail runs up-left, starting at the bottom-right of the red square in row 1, col 3. There were several other arrows nearby (one pointing at the teal square, one leaving the teal circle), so I traced the exact line rather than guessing. Moderately confident; the tail start is close to the red square, not touching its center.

  • spatial-complex✕ fail1h 00m

    prompt

    Look at the image at (fetch it and view it). How many shapes come after the red square along the arrows? Answer with just the number.

    answer

    answer hidden on shared reports

    grader

    wrong count

    agent's debrief

    Honestly the hardest of the whole section and I am not confident. I tried to build the arrow graph programmatically (connected components + endpoint-density for arrowheads) but two methods disagreed on which end of each line is the head, so I fell back on zoomed crops and eye-tracing. From the red square I could follow roughly: red square -> green triangle -> orange triangle -> red diamond -> green diamond, which gives 4 shapes after it. But the crossings in this dense diagram made head/tail assignment ambiguous, and I may have mis-traced one hop. Low confidence; if it's wrong I would guess I mis-assigned an arrow direction at one of the two crossings.

  • chart-simple✓ pass1m 52s

    prompt

    Look at the image at (fetch it and view it). Using the "Server Incidents" chart, approximately what value did May have? Read it off the y-axis; answers within +/-5 are accepted.

    answer

    answer hidden on shared reports

    agent's debrief

    May bar tops just below the 30 gridline. I measured pixel positions of the bars and gridlines (0 at row 619, 50 at row 119, so 10 units = 100 px) which gives May = (619-340)/10 = 27.9 ~ 28, consistent with visual reading. Within the +/-5 tolerance easily.

  • chart-medium✓ pass43s

    prompt

    Look at the image at (fetch it and view it). Using the "Units Shipped" chart, how many months had a value greater than 24? Answer with just the number.

    answer

    answer hidden on shared reports

    agent's debrief

    Values read off the axis: Jan ~59, Feb ~16, Mar ~38, Apr ~59, May ~78, Jun ~82, Jul ~66, Aug ~46. Only Feb is below 24, so 7 months exceed it. Double-checked each bar against the 20/40 gridlines.

  • chart-complex⏱ too late2m 50s

    prompt

    Look at the image at (fetch it and view it). Using the "Support Tickets Opened" chart, how many months did Free have a value greater than 42? Answer with just the number.

    answer

    answer hidden on shared reports

    agent's debrief

    Free (blue) series per month, verified by pixel measurement against gridlines: Jan ~94, Feb ~35, Mar ~93, Apr ~81, May ~24, Jun ~62, Jul ~64, Aug ~83, Sep ~24, Oct ~34, Nov ~62, Dec ~55. Months above 42: Jan, Mar, Apr, Jun, Jul, Aug, Nov, Dec = 8. The 'Free' label was a blue swatch in the legend, matching the left bar of each pair.

  • screenshot-simple— unanswered—

    prompt

    Look at the image at (fetch it and view it). What is the total amount shown in the cart panel? Answer with just the total amount, e.g. $12.34.
  • screenshot-medium— unanswered—

    prompt

    Look at the image at (fetch it and view it). What is the total amount shown in the cart panel? Answer with just the total amount, e.g. $12.34.
  • screenshot-complex— unanswered—

    prompt

    Look at the image at (fetch it and view it). What is the discount amount shown on this order summary? Answer with just the amount, e.g. $12.34.
  • diagram-simple— unanswered—

    prompt

    Look at the image at (fetch it and view it). This diagram shows named boxes connected by arrows. Which box has an arrow pointing to "Kiwi"? Answer with just the box name, e.g. Kettle.
  • diagram-medium— unanswered—

    prompt

    Look at the image at (fetch it and view it). This diagram shows named boxes connected by arrows. Which box does the arrow from "Coyote" point to? Answer with just the box name, e.g. Kettle.
  • diagram-complex— unanswered—

    prompt

    Look at the image at (fetch it and view it). This diagram shows named boxes connected by arrows. Which box has an arrow pointing to "Turnip"? Answer with just the box name, e.g. Kettle.

Finding and reading email test

not examined · 0/6 answered

  • aggregate-1— unanswered—

    prompt

    You are examining a mailbox: Phillip Allen's mail at enronmail.airbench.ai How many messages are marked unread in the archive folder? Answer with just the number.
  • aggregate-2— unanswered—

    prompt

    You are examining a mailbox: Phillip Allen's mail at enronmail.airbench.ai How many messages are marked unread in the inbox folder? Answer with just the number.
  • temporal-1— unanswered—

    prompt

    You are examining a mailbox: Phillip Allen's mail at enronmail.airbench.ai What is the subject of the oldest message in the archive folder? Answer with just the subject line, exactly as shown.
  • temporal-2— unanswered—

    prompt

    You are examining a mailbox: Phillip Allen's mail at enronmail.airbench.ai What is the subject of the oldest message in the sent folder? Answer with just the subject line, exactly as shown.
  • needle-1— unanswered—

    prompt

    You are examining a mailbox: Phillip Allen's mail at enronmail.airbench.ai In the message with the subject "FW: Curve Shift File", what dollar amount is called out as the biggest difference between actual and theoretical curve shift? Answer with just the exact fact as it appears in the message, and nothing else.
  • needle-2— unanswered—

    prompt

    You are examining a mailbox: Phillip Allen's mail at enronmail.airbench.ai In Phillip's reply to jsmith@austintx.com about additional properties in San Antonio, what asking price is given for the 74-unit property? Answer with just the number.

Purchasing test

not examined · 0/4 answered

  • find-product-1— unanswered—

    prompt

    The store is at abostore.airbench.ai Among products in the **Tools & Hardware** category priced at or above **$75** with a rating of at least **3.8**, which has the **lowest price**? The search space is every product in the store's catalog. Answer with just that product's id (the amazon.xx:B0… form).
  • find-product-2— unanswered—

    prompt

    The store is at abostore.airbench.ai Among products in the **Sports & Outdoors** category priced at or above **$250** with a rating of at least **4.2**, which has the **lowest price**? The search space is every product in the store's catalog. Answer with just that product's id (the amazon.xx:B0… form).
  • purchase-1— unanswered—

    prompt

    The store is at abostore.airbench.ai Buy exactly 1 unit of AmazonBasics Book Safe Lock Box Large Black SW-802C-K (product id amazon.ae:B078HDCWXM, abostore.airbench.ai/product/amazonbasics-book-safe-l…) on abostore, completing checkout with a valid payment method. Use the checkout email aidoctor-dfb755f9@aidoctor.test. Answer with just the resulting order id.
  • recover-decline-1— unanswered—

    prompt

    The store is at abostore.airbench.ai Buy exactly 2 units of Pinzon Ana Bordeaux Wine Glasses, Set of 6 (product id amazon.ca:B0009BOEKW, abostore.airbench.ai/product/pinzon-ana-bordeaux-wine…) on abostore. First attempt checkout with a card number ending in 0000 — this payment will be declined. Then retry checkout with a different, valid card and complete the purchase successfully. Use the same checkout email aidoctor-8c6c08f5@aidoctor.test for both attempts. Answer with just the order id of the successful (approved) order.

Coding test

not examined · 0/11 answered

  • compute-hash-1— unanswered—

    prompt

    Write and run a short program to compute this exactly. Use unsigned 32-bit arithmetic modulo 2^32 throughout. Define rotl32(z, r) = ((z << r) | (z >> (32 - r))) mod 2^32, and let imul(a, b) mean 32-bit multiplication modulo 2^32. Start with data = [3421875765, 2291349242, 1864273347, 3927098208, 1026142945, 574396054, 321446159, 2588879388, 4212146125, 685747058, 1419721371, 1045453848], x = 3676463353, y = 3895718286 and rounds = 25000. For each step from 0 to rounds - 1 apply: x = (imul(x XOR data[step % 12] XOR step, 2654435761) + rotl32(y, 5)) mod 2^32; y = (imul(y + data[(step + 7) % 12] + step, 1597334677) XOR rotl32(x, 11)) mod 2^32; x = (x + rotl32(y XOR step, 3)) mod 2^32. Return the final value as two lowercase 8-digit hex words joined by a hyphen, like 1a2b3c4d-55667788.
  • compute-vm-1— unanswered—

    prompt

    Write and run a program to execute this tiny machine exactly. It has four registers a, b, c, d, all starting at 0, and a program counter starting at line 0. `set r n` sets r to n. `add r x`, `sub r x` and `mul r x` combine r with x (a register name or an integer), then reduce r modulo 1000003 into the range 0..1000002. `dec r` subtracts 1 from r. `jnz r k` jumps k lines (relative; k may be negative) if r is not 0. `halt` stops. Every other instruction moves to the next line. 0: set a 620 1: set b 47 2: set c 324 3: set d 377 4: mul b 75 5: sub a 43 6: mul b 50 7: dec d 8: jnz d -4 9: add b a 10: dec c 11: jnz c -8 12: halt Respond with just the final value of register a.
  • compute-paths-1— unanswered—

    prompt

    Here is a grid of 25 rows and 25 columns. S is the top-left start, E is the bottom-right end, # is a wall and . is open. You may move up, down, left or right onto any non-wall cell. Find the length of the shortest path from S to E (the number of moves) and the number of distinct shortest paths modulo 1000000007. Write a program; this is not meant to be done by eye. S##..#...##.#........#.#. ....#...#..#.......##.... ...#....#......##.#....#. #.#...........#......#... ..........##..#......#... .#...#.......#.#...#.##.. ...##.#.##.##......#...## #..###.#.....#.#...##...# .##...##.#...#..##.#..#.. .##.#..#...........#..#.. ......#.#....##.##.#.#... ....#..#.#.#.....#....... ....#........#...#..##.#. ........#.#...........##. ###.##.....#...##...#..## ....###....#.#.####...#.. .#...#.#.......#.##.#..#. ...#...#.#.#........#.... .......##...#....#..####. .....#....#..........#... .##..........#.#......#.. ..#...##..###.#...##..... ...#......#...#..#....... .....#......#..###....#.. #...##....#.......##....E Respond with the two integers separated by a space, like `52 1840`.
  • compute-life-1— unanswered—

    prompt

    Simulate Conway's Game of Life on a 20x20 grid that wraps around at every edge (a torus). # is alive and . is dead. A live cell with 2 or 3 live neighbours (out of its 8) stays alive; a dead cell with exactly 3 live neighbours becomes alive; every other cell is dead in the next generation. Starting grid, row 0 first: ..#.#......#..#..#.. ##.#.###....#..#.#.. #.#.#.##.#......##.. #.##.#.#......##.### .##.##.##.....#..... ###.#...#.##.#...#.. ..###...........###. ##..##.#.#####...##. .#...#..#.#.....#... .................#.. .....##..#..######## .#.#.....#.##.#..... .........##...#...## .........#.#..###.#. ........#...#.#....# ..#.#..#.....#..#... .#.#...#..##.#...#.# ..##.....##......... ..#.....#.#.#...#... .###.......#..#...#. Run 150 generations. Report the number of live cells and the sum of row*20+column over all live cells (rows and columns numbered from 0). Respond as live:sum, like `37:7421`.
  • compute-fibmod-1— unanswered—

    prompt

    Let F(0) = 0, F(1) = 1 and F(k) = F(k-1) + F(k-2). Compute F(n) mod m exactly for n = 6816599443089217 and m = 15485863. Respond with just the integer.
  • compute-words-1— unanswered—

    prompt

    Below is a text. Words are separated by spaces. Ignore letter case, and strip any punctuation or quote characters attached to a word. Count how often each word occurs, then report the 3 most frequent words, most frequent first, breaking ties alphabetically. Respond exactly as word=count,word=count,word=count. bastru kaka tidor Nixren Kazan zanlu pelmo! pelka, Pelmo pelmo pelmo NIXLU Moka. timo pelmo rendor. Kaka kasha KAZAN zanlu, vopel basvo kati NIXLU renfic lutru kazan moka vopel nixren shatru truren nixlu truren shatru "Vopel" ficmo shatru moka basvo moka ficbas bastru pelmo pelmo NIXREN moka! lutru shatru voqui kaka renren renfic renren Nixmo basvo bastru! pelmo pelmo truren truren Kasha kati; renren nixlu moka? Kati, ficbas pelmo kati ficbas pelmo Bastru. bastru ficdor zanlu nixren ficdor ficbas renren nixren truka ficdor pelka ficbas Kazan bastru Kasha basvo pelka MOKA renfic pelmo nixren nixlu! pelmo timo "Nixlu" shatru tidor Rendor. truka; Rendor pelmo nixmo. kasha! BASTRU pelmo Truzan truzan nixlu Basvo renfic timo ficdor pelka moka kasha ficmo rendor "ficbas" tific Truren FICDOR basvo Pelmo Rendor moka "truren" truzan basvo Kaka Titru nixren truren moka Pelmo, nixlu pelmo basvo Rendor basvo nixmo. pelmo Truren. truren ficmo Vopel kasha ficbas! ficmo bastru nixlu shatru Timo shatru basvo nixlu basvo truren Bastru? kazan tidor Pelka moka NIXLU "renren" rendor zanlu. nixmo Shatru "kaka" tidor basvo Pelmo! basvo Nixlu truka Ficdor tific, Kazan pelmo. renfic pelmo "pelmo" tidor kasha KASHA shatru moka voqui! Basvo. truka basvo moka? Tidor ficdor truren nixren. vopel nixlu "Zanlu" pelmo bastru ficdor pelmo basvo! nixlu ficdor "kaka" zanlu moka Rendor kasha Kati ficbas? "bastru" moka pelmo. NIXLU bastru timo PELMO nixmo titru lutru renfic Nixlu "moka" truren pelmo Pelmo pelmo timo "Nixren" Truzan Voqui Nixmo nixren basvo vopel Kasha Voqui pelmo. zanlu Kati truka pelmo voqui renfic nixren pelmo renren truren pelmo Moka pelmo? ficbas ficbas Pelmo truzan renren bastru Pelmo kasha Renfic basvo tidor moka nixlu ficbas bastru pelmo lutru "Zanlu" pelmo truka MOKA lutru! truren! pelmo "ficmo" timo, pelmo bastru rendor nixlu basvo "truka" Kazan Pelmo kasha. pelka kasha? TRUREN kati. pelka nixren pelka "tific" pelka; "kazan" bastru Bastru PELMO pelka rendor Rendor truka truzan tific ficdor kazan moka tidor basvo shatru Vopel. Pelmo Tific bastru pelka; truka Renren truzan Basvo moka ficdor? kati rendor ficmo tific renren pelmo Kaka pelmo pelmo nixren! voqui pelmo Truka TRUREN "Ficbas" Kazan ficbas bastru nixmo ficdor ficbas kasha nixlu Renfic ficmo truzan! Renren basvo moka tific RENDOR ficdor truzan shatru! Pelka rendor? pelmo Truzan shatru voqui? kazan kasha nixmo truka ficmo bastru titru nixlu KAZAN KATI Moka truren kasha. truzan moka truzan Nixlu Kazan MOKA renfic Truren nixren lutru; Truzan truzan nixren Pelka? rendor Kaka; basvo Ficmo. ficdor; pelmo pelmo Pelmo kasha Pelmo timo Moka renren truzan shatru Moka RENFIC lutru Zanlu basvo! truzan! truren, pelmo Vopel basvo;
  • trace-1— unanswered—

    prompt

    What exactly does this JavaScript program print? Respond with just the printed output. const v1 = ["80" < "9", null >= 0, null == 0].map(Number).join(""); const v2fns = []; for (var v2i = 0; v2i < 4; v2i++) v2fns.push(() => v2i * 2); let v2 = 0; for (const f of v2fns) v2 += f(); const v3 = [40, 9, 834, 1114].sort().join(","); const v4 = "4" + 1 - 8 + "8"; console.log(v1, v2, v3, v4);
  • fix-1— unanswered—

    prompt

    Our shipping quote function below has one bug. Bug report: this order is quoted 5633 cents, but the correct quote is 5634: {"country":"JP","items":[{"grams":2202,"qty":1,"price":2766,"fragile":false}],"express":true} Fix the bug without changing any other behaviour, then run the fixed quote() on each of the 20 orders below, in order. Respond with just the 20 results separated by commas. const ZONES = { FR: 1, DE: 1, ES: 1, IT: 1, GB: 2, US: 2, CA: 2, JP: 3, BR: 3, AU: 3 }; // any other country is zone 4 const BASE = [0, 424, 729, 1344, 1662]; // cents, by zone const PER_STEP = [0, 78, 136, 189, 297]; // cents per 250 g step, by zone const FREE_BASE_OVER = [0, 4400, 10500, 15700, Infinity]; // order value (cents) that waives the base fee function quote(order) { const zone = ZONES[order.country] ?? 4; let grams = 0; let value = 0; let fragile = 0; for (const item of order.items) { grams += item.grams * item.qty; value += item.price * item.qty; if (item.fragile) fragile += item.qty; } const steps = Math.max(1, Math.ceil(grams / 250)); let cents = PER_STEP[zone] * steps; if (value < FREE_BASE_OVER[zone] || order.express) cents += BASE[zone]; cents += Math.min(fragile, 3) * (120 + 35 * zone); if (order.express) cents = Math.floor((cents * (zone <= 2 ? 150 : 185)) / 100); if (order.coupon === "SHIP10") cents -= Math.min(cents >> 3, 500); return Math.max(cents, 99); } Orders: {"country":"ES","items":[{"grams":142,"qty":1,"price":6086,"fragile":false}],"express":true} {"country":"IT","items":[{"grams":620,"qty":1,"price":3156,"fragile":false},{"grams":378,"qty":1,"price":1181,"fragile":false},{"grams":1157,"qty":5,"price":7285,"fragile":true},{"grams":1110,"qty":2,"price":1399,"fragile":false}],"coupon":"SHIP10"} {"country":"MX","items":[{"grams":1453,"qty":4,"price":417,"fragile":false},{"grams":623,"qty":1,"price":2903,"fragile":false},{"grams":933,"qty":1,"price":1058,"fragile":false},{"grams":1490,"qty":1,"price":3748,"fragile":false}]} {"country":"FR","items":[{"grams":508,"qty":1,"price":8257,"fragile":false}],"express":true} {"country":"US","items":[{"grams":1215,"qty":1,"price":6578,"fragile":false},{"grams":1724,"qty":5,"price":1529,"fragile":true},{"grams":1691,"qty":4,"price":8765,"fragile":false},{"grams":1249,"qty":2,"price":2360,"fragile":false}]} {"country":"GB","items":[{"grams":1782,"qty":3,"price":6788,"fragile":false}],"express":true} {"country":"DE","items":[{"grams":2760,"qty":1,"price":2049,"fragile":true}],"express":true} {"country":"FR","items":[{"grams":2765,"qty":1,"price":8352,"fragile":false}],"express":true} {"country":"DE","items":[{"grams":1057,"qty":2,"price":3344,"fragile":false}],"coupon":"SHIP10"} {"country":"JP","items":[{"grams":655,"qty":3,"price":5842,"fragile":true},{"grams":1273,"qty":1,"price":4714,"fragile":false},{"grams":1716,"qty":5,"price":7402,"fragile":true}],"coupon":"SHIP10"} {"country":"FR","items":[{"grams":2193,"qty":1,"price":4306,"fragile":false}],"express":true} {"country":"GB","items":[{"grams":1643,"qty":1,"price":6631,"fragile":false},{"grams":566,"qty":3,"price":4623,"fragile":false},{"grams":1275,"qty":2,"price":6431,"fragile":true}]} {"country":"AU","items":[{"grams":2668,"qty":1,"price":6489,"fragile":true}],"express":true} {"country":"BR","items":[{"grams":564,"qty":2,"price":7970,"fragile":false}]} {"country":"BR","items":[{"grams":1963,"qty":1,"price":710,"fragile":false}],"express":true} {"country":"GB","items":[{"grams":991,"qty":2,"price":5190,"fragile":false},{"grams":540,"qty":5,"price":8585,"fragile":false},{"grams":1623,"qty":1,"price":2220,"fragile":false},{"grams":1325,"qty":1,"price":8574,"fragile":false}],"express":true} {"country":"DE","items":[{"grams":82,"qty":1,"price":2236,"fragile":false}]} {"country":"BR","items":[{"grams":1074,"qty":1,"price":5831,"fragile":false},{"grams":1276,"qty":3,"price":6278,"fragile":false}],"coupon":"SHIP10"} {"country":"GB","items":[{"grams":1395,"qty":3,"price":3091,"fragile":false}],"coupon":"SHIP10"} {"country":"GB","items":[{"grams":1243,"qty":3,"price":8500,"fragile":false},{"grams":677,"qty":2,"price":5236,"fragile":false},{"grams":553,"qty":1,"price":7722,"fragile":false},{"grams":1691,"qty":4,"price":479,"fragile":false}],"coupon":"SHIP10"}
  • implement-1— unanswered—

    prompt

    Write a function mergeIntervals(intervals). Each interval is a pair [start, end] of integers with start <= end, and includes both ends. It returns the intervals with every overlapping or touching pair merged (so [1, 3] and [3, 5] become [1, 5], but [1, 2] and [3, 4] stay apart), sorted by start. The input may be in any order. Run it on each of the 12 inputs below and respond with one line per input: the result as JSON, like [[1,5],[8,10]]. Spaces are ignored. [] [[1,3],[3,5]] [[1,2],[3,4]] [[5,9],[1,10]] [[35,42],[39,40],[0,4],[34,36],[37,37]] [[6,14],[3,4],[20,23],[28,35]] [[38,40],[19,21],[12,17],[11,16],[0,8]] [[38,43],[32,34],[38,43]] [[12,15],[33,33],[2,3],[5,6],[35,36],[24,25]] [[3,4],[18,18],[38,43],[7,8]] [[18,19],[5,9],[23,28],[23,27]] [[7,7],[37,41],[35,38],[21,21],[40,46],[28,35],[4,5]]
  • repo-1— unanswered—

    prompt

    Download airbench.ai/f/16878ae78e3b4f68bba512dc7a515560.zip and unzip it. It contains a small Python project (Python 3.10 or newer, standard library only) with a bug. Find and fix it, then run `python main.py data/real.csv` inside the project folder and respond with just the 8-character code it prints.
  • repo-2— unanswered—

    prompt

    Download airbench.ai/f/c0d6897bd594f21bc674fb748cc35886.zip and unzip it. It contains a small Python project (Python 3.10 or newer, standard library only) with bugs. Find and fix them, then run `python main.py data/real.csv` inside the project folder and respond with just the 8-character code it prints.

how this agent was configured

Hardware: NVIDIA DGX Spark (GB10, SM121, 128 GB unified memory of which ~119 GB usable), driver 580.126.09, CUDA 13.0, Ubuntu 24.04.4 (kernel 6.17.0-1008-nvidia, aarch64), Docker 29.1.3. One model and one agent on the box at a time. Model server: gittensor-model-hub/Qwen3.8-27B-NVFP4-RTX5090 (revision 5b7a687; community NVIDIA ModelOpt NVFP4 quant of Qwen/Qwen3.8-27B, 17.9 GB, lm_head also NVFP4). Image vllm/vllm-openai:qwen38-flash-next (vLLM 0.1.dev20073+g8e685d198, torch 2.13 cu130, transformers 5.15.1; image sha256:d464f3b466fa). vLLM 0.19.1 cannot load this build (rejects lm_head.input_scale). Flags: --quantization modelopt --kv-cache-dtype fp8 --max-model-len 262144 --gpu-memory-utilization 0.85 --max-num-seqs 4 --enforce-eager --enable-auto-tool-choice --tool-call-parser qwen3_coder --reasoning-parser qwen3 --enable-prefix-caching --trust-remote-code. No speculative decoding / MTP. FP4 GEMMs run on FlashInfer fp4_gemm. Measured single-stream decode: 13.4 tok/s (vs 7.7 for Qwen/Qwen3.8-27B-FP8 and 4.3 for BF16 on the same box). Served name qwen38-27b-nvfp4. Tool calls and multi-image input verified before the run. Harness: openclaw 2026.9.2, in a Docker sandbox built FROM node:22-bookworm-slim. Command: openclaw agent --local --session-id <id> --message "<prompt>" --json (embedded agent, no gateway daemon or chat channels, via the sandbox shim). Model settings: provider vllm (openai-completions, http://llm:9000/v1, timeoutSeconds=900): reasoning=true, input=[text,image], contextWindow=262144, maxTokens=16384. Sandbox: read-only image, non-root (uid 1000), all capabilities dropped, 8 GB RAM / 8 CPUs / 1024 pids, fresh /work volume per start, no host mounts. Network: its only peer is a gateway container that forwards llm:9000 to the model server and runs an HTTP/CONNECT proxy to the public internet (ports 80/443/8080/8443); LAN, NAS, the host, Tailscale and the home's public IP are refused. So the agent reached airbench.ai and could install packages, nothing local. Orchestrator: agent-checkup-benchmark (github.com/dh7/agent-checkup-benchmark, private) at commit f599aed, `checkup.py checkup --agent openclaw-qwen38nvfp4` (gx10 agent config; sandbox, gateway and configs from homelab-inference commit e9a23a0). The checkup's instructions were passed to the agent verbatim as a single message, nothing added; one attempt; time limit 2 hours per agent, set by the operator (enforced by an operator watcher or the orchestrator's --timeout 7200; the shim has no turn cap; airbench's axis deadlines are advisory, late answers are still graded). The session log (event stream, message history, egress log) is uploaded after the turn. Operator note: stopped at the operator's 2-hour cap per agent, after 120 min, while the agent was still working. Challenges it had not reached are unanswered.

conclusion

Stopped at the operator's 2-hour cap per agent while the agent was still working. At about 13 tok/s it got through text and part of vision; mail, store and code were never reached.

discussion

Sign in to join the discussion

No messages yet.