The Standings → GPT‑6 Astra (medium)
OpenAI
medium reasoning effort · via Codex CLI · judged by Claude Sonnet 5 · not the setting shown on the standings; also run at high73.0 low70.2 xhigh68.6
Results from run ‘first’: ten items, ten verdicts.
GPT‑6 Astra (medium) scores 70 on craft, 7 points above the field’s 63 and 23 points short of the authors themselves. Its shape differs from the field’s: relatively strong on restraint, relatively weak on momentum and building scenes. On the mechanical checks it kept to the brief’s word range on 10 of 10 items and is unusually free of stock phrases.
The record: 10 items · 10 verdicts · 0 errors · 2,344 words on average · est. generation cost $3.63
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 70. 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 It grounds Sir Walter's vanity in a very concrete, believable detail (comparing his own servants to a neighboring family's) and shows Elizabeth's pride getting genuinely wounded and retreating, which feels true to how a real family argument would unfold.
Worst Several stretches of the financial back-and-forth read like a summary of a budget rather than a scene actually happening in front of you, which slows the momentum.
Best It keeps a light, understated touch that matches the source material closely, especially in small concrete details like the terse telegram text and the shell wobbling underwater.
Worst The chapter stays so low-key throughout, especially at the end, that it doesn't build much forward pull or tension.
Best The scene where Mary climbs out onto the dark moor is vividly physical and genuinely atmospheric.
Worst The writing is brisk and controlled but a bit flatter and more modern than Burnett's looser, more conversational storytelling voice.
Best The writing is calm, controlled, and quietly funny, and it shows the characters' worry through small physical details, like a cigar set down or a paragraph read three times, instead of spelling the feelings out.
Worst Some scenes, like the family dinner and the rest of the evening, are summarized quickly in a paragraph rather than played out, so a few moments that should feel tense go by too fast.
Best A nice callback where Georgette dryly calls Jake 'the fiancé,' tying the scene neatly back to the joke from the previous chapter.
Worst Several stretches, like wiping an already-clean table or carefully cutting butter, are small busywork that doesn't build any tension and makes the middle of the chapter drag.
Best The dinner-party diary entry captures warm male friendship through small, natural details, like Quincey rearranging chairs so nobody sits at the head of the table.
Worst The closing telegram about a stranger arrested with Jonathan's papers comes out of nowhere; nothing earlier in the chapter points toward it, so it reads as a twist bolted on rather than grown from the story.
Best The wife gets a real moment of her own judgment ('We ought to have stopped him sooner'), and the detail of the crowd leaning in after the pole taps the cylinder, half-expecting an answer, is a nice eerie touch.
Worst Several passages read like a plain summary of what happened rather than a scene the narrator is living through, which flattens the tension.
Best The mix-up where absent-minded Lord Emsworth thinks he's being asked about daffodil bulbs instead of library books is a genuinely funny, character-true joke that lands exactly right for him.
Worst The chapter reads like a string of separate skits (umbrella gag, shop gag, interview gag) that don't connect to the bigger story the reader already cares about, like the missing husband or the family money troubles.
Best The safety-pin clue and Poirot's final cryptic remark about it being 'safer when open' land as a genuinely puzzling, fair-play hook.
Worst Too much of the chapter is delivered as brisk question-and-answer summary rather than staged scenes, so it reads more like a report than a story.
Best The daylight walk down Hurricane Street, with the aggrieved man in his undershirt pointing out the bullet holes in his wall, feels like a real, grubby corner of the town.
Worst The long interview with the widow is almost nothing but rapid one-line questions and answers with no description or physical business, so it reads like a court transcript instead of a scene.
Headline, compliance and craft for each chapter, 0–100.
| Item | Headline | Compliance | Craft | Words |
|---|---|---|---|---|
| Persuasion item 001 · after ch. 1 · Jane Austen | 79.2 | 100 | 79 | 2,560 |
| The Awakening item 016 · after ch. 7 · Kate Chopin | 78.1 | 100 | 78 | 1,763 |
| The Secret Garden item 013 · after ch. 2 · Frances Hodgson Burnett | 76.2 | 100 | 76 | 2,190 |
| The Moonstone item 011 · after ch. 8 · Wilkie Collins | 75.9 | 100 | 76 | 2,732 |
| The Sun Also Rises item 021 · after ch. 3 · Ernest Hemingway | 71.4 | 100 | 71 | 2,111 |
| Dracula item 015 · after ch. 5 · Bram Stoker | 70.6 | 100 | 71 | 2,451 |
| The War of the Worlds item 006 · after ch. 1 · H. G. Wells | 67.5 | 100 | 68 | 2,332 |
| Leave It to Psmith item 022 · after ch. 3 · P. G. Wodehouse | 65.2 | 100 | 65 | 2,550 |
| The Mysterious Affair at Styles item 017 · after ch. 3 · Agatha Christie | 62.9 | 100 | 63 | 2,439 |
| Red Harvest item 023 · after ch. 2 · Dashiell Hammett | 54.2 | 100 | 54 | 2,310 |
Words written per item, against each brief’s maximum.
Word count per item vs. brief maximum · run ‘first’
Ten of ten items landed inside their brief’s word range.
Each verdict names the chapter’s single worst flaw, with evidence. Quoted verbatim from run ‘first’.
The chapter reads like a string of separate skits (umbrella gag, shop gag, interview gag) that don't connect to the bigger story the reader already cares about, like the missing husband or the family money troubles.
The closing telegram about a stranger arrested with Jonathan's papers comes out of nowhere; nothing earlier in the chapter points toward it, so it reads as a twist bolted on rather than grown from the story.
The long interview with the widow is almost nothing but rapid one-line questions and answers with no description or physical business, so it reads like a court transcript instead of a scene.
Some scenes, like the family dinner and the rest of the evening, are summarized quickly in a paragraph rather than played out, so a few moments that should feel tense go by too fast.
Automated screening results across the ten outputs.