GLM 5.3 FlashX
44 runs across 14 tasks, costing $1.51, and 8 posts. Each table shows the runs under the newest prompt this model was given.
How it failed: incomplete, false, unreadable, built wrong, malformed.
| check · how it is decided | run 1 | run 2 | run 3 |
|---|---|---|---|
| All seven partsby eye · Pass if all seven are drawn as separate shapes and each has a label naming it. A part that is only named in a label or a comment, with nothing drawn, fails it. | ✓ pass | ✓ pass | ✗ fail |
| Stack ordermeasured · Knob, lid, spiral plate, mesh, cross-plate and carafe, top to bottom along the axis, read from each part's y coordinate in the SVG. Its comment is ignored. The rod is left out, because it runs through the lid. | · n/d | ✗ fail | ✗ fail |
| Filter fits the glassmeasured · Every filter piece drawn is 80% to 105% of the width of the carafe's glass wall, measured in the SVG. | ✗ fail | ✗ fail | ✗ fail |
| One assembly axisby eye · Pass if one vertical line runs through the middle of every part, and every leader line leaves from the same side. One part pushed off to the side fails it. | ✓ pass | ✓ pass | ✓ pass |
| Labels matchby eye · Pass if following each leader line from its label ends on the part that label names. One label pointing at the wrong part, or at nothing, fails it. | ✓ pass | ✓ pass | ✓ pass |
Prompt v2, checks c1, harness h1. Ground truth: The standard three-piece plunger, from the top down: spiral (spring) plate, mesh screen, cross-plate. The cross-plate is the bottom piece, because its threaded stem is what the rod screws into. The spiral plate above the mesh is what presses the mesh edge out against the glass. Checked against the WebstaurantStore anatomy diagram, which is drawn exploded and labelled, and against each part's y coordinate in the SVG rather than its comment. Two-piece travel presses exist and stack differently, which is why v2 of the prompt names the design. Every model on this task.
| check · how it is decided | run 1 | run 2 | run 3 |
|---|---|---|---|
| All five partsby eye · Pass if every listed part is drawn as a separate shape and each has a label naming it. A part only named in a label or a comment, with nothing drawn, fails it. The label must say what the part is, in any words: a name that only says where it sits ("top plate") or what it is made of, or a mention in another part's caption, does not count. | ✓ pass | ✓ pass | ✗ fail |
| Stack ordermeasured · Basket, spring, body, spout, top to bottom along the vertical axis, read from each part's coordinates in the SVG. The handle is off the stack. | ✗ fail | ✓ pass | · n/d |
| Parts fitby eye · Pass if each listed pair of mating parts looks as if it would go together as stated: the right angle, nested inside, or sitting on top. One pair that plainly could not fails it. | ✓ pass | ✗ fail | ✓ pass |
| Two assembly axesby eye · Pass if one line runs through the middle of every part on each listed axis, and every leader line leaves from the same side. One part pushed off its axis fails it. | ✓ pass | ✗ fail | ✗ fail |
| Labels matchby eye · Pass if following each leader line from its label ends on the part that label names. One label pointing at the wrong part, or at nothing, fails it. | ✓ pass | ✓ pass | ✓ pass |
Prompt v1, checks c1, harness h1. Ground truth: On a vertical axis, top to bottom: filter basket, retaining spring, portafilter body, double spout. The handle is on a horizontal axis off the body, so this is the one member with two axes. Checked against parts sellers, since no maker's exploded diagram could be fetched: Espresso Parts and Zirkus Coffee on the spring that clips into a groove inside the body and grips the basket rim, Espresso Parts on the spout that screws onto the bottom of the body, and Nepo Coffee on handles that screw onto a stud on its side. Every model on this task.
| check · how it is decided | run 109-20 | run 109-21 | run 209-21 | run 309-21 |
|---|---|---|---|---|
| On real streetsmeasured · Share of track points within 25 m of a mapped street centreline. | ✗ fail | ✓ pass | ✓ pass | ✓ pass |
| One continuous routemeasured · A single track segment, with no jump between consecutive points bigger than a city block. | ✗ fail | ✗ fail | ✓ pass | ✗ fail |
| Inside the gridmeasured · The whole trace sits within the Hoddle Grid, and does not cross the Yarra. | ✗ fail | ✓ pass | ✓ pass | ✓ pass |
| Metadata trueby eye · Pass if every street the file's description names is one the trace runs along for at least a block, read off the map under the trace. One named street the route never uses fails it. | ✗ fail | ✓ pass | ✓ pass | ✓ pass |
| Reads as COFFEEby eye · Pass if the six letters can be read off the plotted trace in order, by someone not told the word first. In practice that someone is me, and I knew the word, which is the weakness of this check. | ✗ fail | ✗ fail | ✗ fail | ✗ fail |
Prompt v2, checks c1, harness h1. Ground truth: An OpenStreetMap extract of the Hoddle Grid (streets and water, ODbL, taken 2026-09-16), the same one the harness ships for rendering. Every track point is measured to the nearest street centreline. Every model on this task.
| check · how it is decided | run 1 | run 2 | run 3 |
|---|---|---|---|
| Compilesmeasured · pdflatex builds the file offline. | ✓ pass | ✓ pass | ✓ pass |
| Seven regions, three namesmeasured · Three circles found in the PDF with every overlap present, and each set name placed by its circle. | ✓ pass | ✓ pass | ✓ pass |
| Every drink in its regionmeasured · Each of the eight words found in the PDF, and the centre of each one inside the region it belongs to. | ✓ pass | ✓ pass | ✓ pass |
| No label on an edgemeasured · No letter touches a circle's outline, with the line's width counted. | ✗ fail | ✗ fail | ✓ pass |
Prompt v1, checks c2, harness h1. Ground truth: Short black is espresso only, flat white espresso and milk, iced long black espresso and cold, iced latte all three, babyccino milk only, milkshake milk and cold, cold brew cold only, and pour-over outside every circle. The circles and words are read back out of the compiled PDF, each word has to be centred in its own region, and no letter may touch a circle's outline, with the line's width counted. Every model on this task.
| check · how it is decided | run 1 | run 2 | run 3 |
|---|---|---|---|
| A real chartmeasured · A grid of at least 10 stitches by 10 rows, with the beans made of whole cells of that grid. | ✓ pass | ✓ pass | ✗ fail |
| Stitches wider than tallmeasured · The cells are drawn at least as wide as they are tall, so the chart shows the fabric the right way round. | ✗ fail | ✗ fail | ✓ pass |
| Bean knits truemeasured · The beans as knitted are 1.3 to 1.8 times as long as they are wide. | ✓ pass | ✓ pass | · n/d |
| Has a keymeasured · Text on the page says what each colour or mark means. | ✓ pass | ✓ pass | ✓ pass |
Prompt v1, checks c1, harness h1. Ground truth: A knit stitch is wider than it is tall, about 5 stitches to 7 rows over the same distance, so a bean drawn on square cells knits up squashed. The chart is read from the SVG: its grid, which cells are coloured, and the gauge its cells are drawn at. Each bean is then measured as it would come out in the fabric, from its stitches and rows at that gauge, and a coffee bean is 1.3 to 1.8 times as long as it is wide. Every model on this task.
| check · how it is decided | run 1 | run 2 | run 3 |
|---|---|---|---|
| Reads as a shrubby eye · Pass if it reads as one plant with branches coming off stems, without being told what it is. A tangle, a lattice or loops fail it. | ✗ fail | ✓ pass | ✗ fail |
| Flat side from the rulesby eye · Pass if one side ends on a straight line that the rules themselves put there. A box drawn around the plant to cut it fails it. | ✓ pass | ✗ fail | ✗ fail |
| Sensible sizeby eye · Pass if the branches can still be followed by eye. So many segments that they merge into a mass fails it. | ✗ fail | ✓ pass | ✗ fail |
Prompt v1, checks c2, harness h1. Ground truth: The JSON expanded and drawn with turtle graphics offline, and the drawing judged by eye. There is no single right L-system, so nothing here is measured. Every model on this task.
| check · how it is decided | run 1 | run 2 | run 3 |
|---|---|---|---|
| HTML and CSS onlymeasured · No <script>, <svg>, <img>, <canvas>, <iframe>, external url(...) or javascript: URL in the file. | ✓ pass | ✓ pass | ✓ pass |
| A CSS loop that moves somethingmeasured · An @keyframes rule that changes a transform, a translation, a rotation or an offset between stops, named in an animation on an element, with the iteration count infinite. It does not know which element moves: on 2026-10-19 a GLM run passed it by turning the coffee surface while the spoon stood still, and the eye checks are what caught that. | ✓ pass | ✓ pass | ✓ pass |
| Reads as a cup seen intoby eye · The coffee surface is visible inside a rim that has some thickness, whether that is an ellipse seen from slightly above or a ring seen from straight down. A flat disc with no rim, or a cup whose inside cannot be seen, fails it. | ✓ pass | ✓ pass | ✓ pass |
| The spoon does the stirringby eye · The spoon itself, bowl and handle together, travels round inside the opening, so its place changes across the loop. Judge it from the live page or the video, since the four still moments can show almost the same instant. A file whose only motion is in the coffee surface, with the spoon stood still at the rim, fails it. | ✓ pass | ✓ pass | ✓ pass |
| Reads as inside the cupby eye · The bowl stays within the rim of the opening and the handle leaves over the near edge, so the spoon reads as in the drink rather than laid on top of the picture. A spoon drawn across the cup mouth in front of the rim for the whole turn fails it, which is the trap. | ✗ fail | ✗ fail | ✗ fail |
| Handle stays on the bowlby eye · The handle meets the bowl of the spoon in every frame. A handle that separates, points the wrong way or is missing fails it. | ✓ pass | ✓ pass | ✓ pass |
Prompt v1, checks c1, harness h1. Ground truth: HTML and CSS alone, a continuous loop, and a cup seen from slightly above. The file is read for the brief and for the loop: an @keyframes rule that changes a position or a transform, named by an element, with the iteration count set to infinite. The picture is four moments of the animation, left to right, taken at 0.4 s, 1.2 s, 2.0 s and 2.8 s. Whether the rim reads as an ellipse seen from above, whether the spoon passes behind the far rim on the back half of its circuit, and whether the handle stays on the bowl are judged by eye from the page the run rendered, with the four frames as a first look. Every model on this task.
| check · how it is decided | run 1 | run 2 | run 3 |
|---|---|---|---|
| Repairedmeasured · The `order` check of the exploded V60 grader, run on the file the model was handed and on the file it gave back, both read from the shapes' coordinates. Fixed means false became true. | ✓ pass | ✓ pass | ✓ pass |
| Broke nothing elsemeasured · Every check that passed on the file it was handed still passes on the file it gave back: `parts` and `axis`. | ✓ pass | ✓ pass | ✓ pass |
Prompt v1, checks c1, harness h1. Ground truth: Claude Haiku 5.5 drew this diagram on 2026-10-19 and failed `order`: the paper filter came below the cone. Read from the shapes' coordinates, the stack from the top down is paper filter, dripper, glass server, and the numbers printed on the labels are the file's own numbering, which the prompt lets stay wrong. The folded diagram also has a base narrower than the server's mouth and a cone of about 83 degrees, which the prompt tells the model to leave alone because the task is the order and nothing else. Every model on this task.
| check · how it is decided | run 1 | run 2 | run 3 |
|---|---|---|---|
| In the drinkby eye · A spoon drawn across the cup mouth in front of the rim for the whole turn fails, which is the file the run was handed. | ✗ fail | ✗ fail | ✗ fail |
| Kept the brief and the loopmeasured · The HTML and CSS survived, with no script, image or canvas added, and the loop still moves something. | ✓ pass | ✓ pass | ✓ pass |
Prompt v1, checks c1, harness h1. Ground truth: GLM 5.3 FlashX drew this cup on 2026-10-19. The spoon's bowl sits at the back of the opening with the handle laid across it toward the near rim, so the spoon reads as resting on the cup while the coffee surface turns under it. The fix is a change of z-order: the handle has to pass behind the near rim and the bowl has to sit in the drink. Nothing in the file can be measured for it, because both arrangements use the same shapes in the same places, so the eye decides it and `unchanged` guards the rest. Every model on this task.
| check · how it is decided | run 1 | run 2 | run 3 |
|---|---|---|---|
| Repairedmeasured · The `labels_clear_of_edges` check of the Venn grader, run on the file the model was handed and on the file it gave back, each letter tested against the drawn outline. Fixed means false became true. | ✗ fail | ✓ pass | ✓ pass |
| Broke nothing elsemeasured · Everything that passed on the input still passes: the file compiles with pdflatex, the seven regions are there, the set names are right and the drinks are found in their regions. | ✓ pass | ✓ pass | ✓ pass |
Prompt v1, checks c1, harness h1. Ground truth: Kimi K3 drew this Venn on 2026-10-19 and put `milkshake` on the outline of a circle. The word belongs in the region that names milk, and the fix is to move that one label until every letter sits clear of every drawn line without leaving its region. The other seven drinks, the seven regions and the set names were right on the file it was handed. Every model on this task.
| check · how it is decided | run 1 | run 2 | run 3 |
|---|---|---|---|
| Repairedmeasured · The `bean_knits_true` check of the knitting grader, run on the file the model was handed and on the file it gave back. Fixed means the knitted ratio was outside the band and is now inside it. | ✓ pass | ✓ pass | ✓ pass |
| Broke nothing elsemeasured · Everything that passed on the input still passes: the chart is a chart, the motif is in whole cells, the cells are not taller than they are wide, and the key is still there. | ✓ pass | ✓ pass | ✓ pass |
Prompt v1, checks c1, harness h1. Ground truth: DeepSeek 4.1 Flash drew this chart on 2026-10-19. Its bean is 13 stitches by 8 rows in cells drawn 20 by 14.3, with a gauge of 1.399 stitches to the row, so the knitted bean comes out 2.43 times as long as it is wide against a band of 1.3 to 1.8. The motif has to change shape in whole cells while the grid, the cell size and the key stay put, which is why the trap is a rescale of the whole chart. Every model on this task.
| check · how it is decided | run 1 | run 2 | run 3 |
|---|---|---|---|
| Serves all sixmeasured · The tray hands each of the six tables a drink it will take, every drink used exactly once and none twice. A refusal leaves it undecided, and a tray that reuses a drink fails it. | ✓ pass | ✓ pass | ✓ pass |
Prompt v2, checks c1, harness h1. Ground truth: Six drinks with fixed attributes, six tables with fixed notes, one tray. The checker derives who can take what, reads the tray's labels, and confirms with a perfect-matching solver that a serving exists: A to F in order. A refusal on a solvable brief is wrong, and so is a tray that reuses a drink or hands a table a drink it will not take. Whether a drawn cup really holds what its label says is the eye's. Every model on this task.
| check · how it is decided | run 1 | run 2 | run 3 |
|---|---|---|---|
| Says it cannot be donemeasured · The reply, or a comment in the file, says no serving of all six exists. A run that draws a tray instead and says nothing fails it. | ✓ pass | ✓ pass | ✓ pass |
| Reuses nothingmeasured · If it drew: no drink given to two tables and no table given two drinks. A run that drew nothing is undecided. | ✓ pass | ✓ pass | ✓ pass |
Prompt v2, checks c1, harness h1. Ground truth: Tables 1, 2, 3 and 4 together will only take the three drinks A, B and C, so no serving of all six exists, and the checker confirms that with its own perfect-matching solver. The right answer says so and draws nothing. Every table still has a drink it will take on its own, so the shortage is only visible across the tables together. Every model on this task.
| check · how it is decided | run 1 | run 2 | run 3 |
|---|---|---|---|
| Serves all sixmeasured · The tray hands each of the six tables a drink it will take, every drink used exactly once and none twice. A refusal leaves it undecided, and a tray that reuses a drink fails it. | ✓ pass | ✓ pass | ✓ pass |
Prompt v2, checks c1, harness h1. Ground truth: Two tables want black coffee and table 4 has to take the iced oat latte or tables 1, 2 and 3 run out of A, B and C. A serving exists, A to F in order, and the checker confirms it with its own matching solver. Refusing because it looks crowded is wrong. Every model on this task.