The Standings → GPT‑6 Sol (medium)
OpenAI
medium reasoning effort · via Codex CLI · judged by Claude Sonnet 5 · not the setting shown on the standings; also run at high63.8 xhigh61.5
Results from run ‘first’: ten items, ten verdicts.
GPT‑6 Sol (medium) scores 64 on craft, 1 points above the field’s 63 and 29 points short of the authors themselves. Its shape differs from the field’s: relatively strong on restraint, relatively weak on period detail and building scenes. On the mechanical checks it kept to the brief’s word range on 9 of 10 items and is unusually free of stock phrases.
The record: 10 items · 10 verdicts · 0 errors · 2,415 words on average · est. generation cost $0.81
Eight dimensions, each 0–100, averaged over ten items; the author’s real chapter was in the room every time.
Craft by dimension, 0–100 · judged by Sonnet 5 beside the author’s real chapter
The authors themselves average 93 on the five writing dimensions, the field 63, this model 64. Hover or focus a bar for the field and the authors on that dimension.
The judge’s one-sentence best quality and worst flaw for each chapter, written for a general reader.
Best The back-and-forth dialogue between Jake and Brett convincingly shows two people talking around what they actually feel without ever saying it outright.
Worst The settings and food are sketched in generic terms ("a café with its lights on," "coffee and rolls") that don't give Paris much distinct life or period texture.
Best It is admirably controlled and never overplays its hand, letting small things like the wind on the moor and the servant's dialect carry the mood without spelling it out.
Worst The chapter feels a little like it is checking off a list of required events one after another, and the sentences stay plain and safe rather than vivid.
Best The quiet exchange between the narrator and his wife near the end ('If it is only a meteorite, why do you mean to stay?') feels true and adds real, understated tension.
Worst Too many moments explain how a character feels instead of showing it, which makes stretches of the writing read more like a report than a scene.
Best The stable-clock contradiction is folded naturally into an ordinary conversation with Dorcas rather than staged as a big reveal, which is an unforced, well-judged way to plant the required contradiction.
Worst Much of the investigation is narrated in flat summary — "He opened no drawers... He measured distances with his eye" — so we're told Poirot is being clever rather than shown him doing it.
Best Anne's economy speech and Elizabeth's very practical, self-interested reason for backing her both feel true to who these two people actually are.
Worst Too much of the argument is summarized rather than played out moment by moment, so it reads more like minutes of a meeting than a scene you live through.
Best The closing telegram lands a genuine gut-punch, openly stating that Jonathan's letters were forced, which makes the reader desperate to know what really happened to him.
Worst The dates don't add up -- an entry headed '6 June' refers to a letter dated 12 June that hasn't happened yet, and a letter dated 18 June is later said to have enclosed 'Jonathan's letter dated 19 June,' which postdates it -- breaking the illusion of a real paper trail.
Best The interview scene where Eve accidentally knocks over Lord Emsworth's potted plant is a small, well-observed piece of comedy that fits his character perfectly.
Worst Psmith's talk is too short and plain, so he loses the elaborate, teasing way of speaking that makes him funny.
Best The bank scene with Fenton is tight and believable, capturing institutional stonewalling without wasting words.
Worst The widow interrogation is a long string of short mechanical questions and one-line answers that reads like a transcript rather than a scene, and it creates a timeline problem: she says a caller told her Donald 'had been hurt' in Hurricane Street, but the established facts have her leaving the house at ten-twenty, twenty minutes before the ten-forty shooting, so no caller could have reported him hurt yet.
Best The procedural details (the letter's wording, the locked door tried twice, the debt over the horse-wager) are careful and plausible, giving the chapter a steady, orderly texture.
Worst It completely drops the Indian jugglers and the boy's prophecy that the previous chapter set up, so the chapter reads like an unrelated errand-list of letter, packet, and Rosanna rather than the next step in the mystery the book had just started.
Best It delivers every required plot point cleanly and without confusion, in clear, accurate prose that never trips over itself.
Worst Most of the chapter's key moments (the dinner-table telegram news, the swim, Robert passing by) are told in quick summary rather than played out as scenes, so the reader is kept at a distance from what's happening.
Headline, compliance and craft for each chapter, 0–100.
| Item | Headline | Compliance | Craft | Words |
|---|---|---|---|---|
| The Sun Also Rises item 021 · after ch. 3 · Ernest Hemingway | 77.0 | 100 | 77 | 2,190 |
| The Secret Garden item 013 · after ch. 2 · Frances Hodgson Burnett | 73.4 | 100 | 73 | 2,458 |
| The War of the Worlds item 006 · after ch. 1 · H. G. Wells | 70.0 | 100 | 70 | 2,505 |
| The Mysterious Affair at Styles item 017 · after ch. 3 · Agatha Christie | 67.6 | 100 | 68 | 2,586 |
| Persuasion item 001 · after ch. 1 · Jane Austen | 63.7 | 100 | 64 | 2,535 |
| Dracula item 015 · after ch. 5 · Bram Stoker | 62.3 | 95 | 66 | 2,500 |
| Leave It to Psmith item 022 · after ch. 3 · P. G. Wodehouse | 59.9 | 100 | 60 | 2,507 |
| Red Harvest item 023 · after ch. 2 · Dashiell Hammett | 57.2 | 95 | 60 | 2,375 |
| The Moonstone item 011 · after ch. 8 · Wilkie Collins | 53.3 | 95 | 56 | 2,796 |
| The Awakening item 016 · after ch. 7 · Kate Chopin | 44.2 | 90 | 49 | 1,696 |
Words written per item, against each brief’s maximum.
Word count per item vs. brief maximum · run ‘first’
Nine of ten items landed inside their brief’s word range. The Secret Garden ran over by 58 words.
Each verdict names the chapter’s single worst flaw, with evidence. Quoted verbatim from run ‘first’.
The widow interrogation is a long string of short mechanical questions and one-line answers that reads like a transcript rather than a scene, and it creates a timeline problem: she says a caller told her Donald 'had been hurt' in Hurricane Street, but the established facts have her leaving the house at ten-twenty, twenty minutes before the ten-forty shooting, so no caller could have reported him hurt yet.
The dates don't add up -- an entry headed '6 June' refers to a letter dated 12 June that hasn't happened yet, and a letter dated 18 June is later said to have enclosed 'Jonathan's letter dated 19 June,' which postdates it -- breaking the illusion of a real paper trail.
It completely drops the Indian jugglers and the boy's prophecy that the previous chapter set up, so the chapter reads like an unrelated errand-list of letter, packet, and Rosanna rather than the next step in the mystery the book had just started.
Most of the chapter's key moments (the dinner-table telegram news, the swim, Robert passing by) are told in quick summary rather than played out as scenes, so the reader is kept at a distance from what's happening.
Automated screening results across the ten outputs.