The Standings → GPT‑5.6 Luna
OpenAI
medium reasoning effort · via Codex CLI · judged by Claude Sonnet 5 · not the setting shown on the standings; also run at xhigh50.3 high48.4
Results from run ‘first’: ten items, ten verdicts.
GPT‑5.6 Luna scores 54 on craft, 9 points below the field’s 63 and 39 points short of the authors themselves. Its shape differs from the field’s: relatively strong on momentum and building scenes, relatively weak on the author’s voice and folding the brief into the story. On the mechanical checks it kept to the brief’s word range on 6 of 10 items.
The record: 10 items · 10 verdicts · 0 errors · 2,583 words on average · est. generation cost $0.08
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 54. 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 dinner scene at Quincey's rooms is a real dramatized scene with lively, in-character dialogue rather than mere summary.
Worst The closing telegram flatly names 'Castle' and 'Count' and has Harker free in Budapest with his own clothes, which both spells out the mystery the chapter was supposed to keep hidden and contradicts the established fact that the Count took his clothes and papers.
Best The scene is cleanly built and paced, with a real back-and-forth argument between the narrator and Ogilvy that escalates naturally to the burn incident.
Worst The dialogue and description read like a competent generic period drama rather than genuinely sounding like Wells's narrator, who normally digresses, reflects, and philosophizes as he tells the story.
Best It is clean and well organized, moving through every required beat without confusion or wasted motion.
Worst The dialogue is so clipped and rapid-fire it reads like a stage script rather than Betteredge's rambling, opinionated voice.
Best It hits every required plot point cleanly and in a sensible order, so the chapter is easy to follow from telegram to swim to the final image.
Worst Feelings are announced instead of dramatized — lines like "the knowledge rested in her like a secret flame" and Robert's remark staying in her mind "tenderly, persistently, like a voice calling from a distance" tell the reader exactly what to feel rather than letting the scene create the feeling.
Best The deadpan interview with Baxter, full of bibliography jargon and Eve's straight-faced nonsense answers, is a nicely dry comic scene.
Worst It quietly swaps in the secretary Baxter for the Earl who was supposed to do the interviewing, and Psmith basically lets Eve's guess about the umbrella stand instead of dodging it the way the real character would.
Best It hits every required beat cleanly and in a logical, easy-to-follow order.
Worst Much of the dialogue is short, quick question-and-answer exchanges that read like a modern script rather than 1911 storytelling prose.
Best It hits every required plot point cleanly and the final image of the darkening room with the question still open lands the required ending well.
Worst Almost every scene is written as a rapid back-and-forth quiz (short question, short answer, short question), which makes the characters feel like they're reciting a script instead of talking like people.
Best The investigation is organized into a clear, brisk sequence of small clues (ashes, torn paper, a maid's contradiction) that keeps the puzzle mechanism easy to follow.
Worst The dialogue is often flat and modern-sounding rather than period Christie — lines like 'That is very precise' or 'I have nothing to conceal' lack Poirot's usual color, French phrasing, and personality.
Best The escalating argument is clear and easy to follow, with each proposal and rejection landing in logical sequence.
Worst It reads like a transcript of an argument, with stage directions between lines of dialogue, rather than the wry, narrator-driven storytelling Austen actually uses.
Best It hits every required plot point cleanly and clearly, so a reader never loses track of what is happening or why.
Worst Almost the entire chapter is made of clipped one- or two-word exchanges strung together, which quickly feels mechanical and repetitive instead of tense or alive.
Headline, compliance and craft for each chapter, 0–100.
| Item | Headline | Compliance | Craft | Words |
|---|---|---|---|---|
| Dracula item 015 · after ch. 5 · Bram Stoker | 60.5 | 90 | 67 | 2,538 |
| The War of the Worlds item 006 · after ch. 1 · H. G. Wells | 57.1 | 100 | 57 | 2,270 |
| The Moonstone item 011 · after ch. 8 · Wilkie Collins | 55.8 | 100 | 56 | 2,789 |
| The Awakening item 016 · after ch. 7 · Kate Chopin | 53.9 | 95 | 57 | 1,901 |
| Leave It to Psmith item 022 · after ch. 3 · P. G. Wodehouse | 50.0 | 85 | 59 | 3,209 |
| The Secret Garden item 013 · after ch. 2 · Frances Hodgson Burnett | 49.5 | 90 | 55 | 2,699 |
| Red Harvest item 023 · after ch. 2 · Dashiell Hammett | 46.4 | 95 | 49 | 2,365 |
| The Mysterious Affair at Styles item 017 · after ch. 3 · Agatha Christie | 46.3 | 85 | 55 | 2,457 |
| Persuasion item 001 · after ch. 1 · Jane Austen | 45.0 | 95 | 47 | 3,018 |
| The Sun Also Rises item 021 · after ch. 3 · Ernest Hemingway | 34.6 | 95 | 36 | 2,588 |
Words written per item, against each brief’s maximum.
Word count per item vs. brief maximum · run ‘first’
Six of ten items landed inside their brief’s word range. The Awakening and Leave It to Psmith and The Secret Garden and Persuasion ran over by 101 and 409 and 299 and 418 words respectively.
Each verdict names the chapter’s single worst flaw, with evidence. Quoted verbatim from run ‘first’.
Feelings are announced instead of dramatized — lines like "the knowledge rested in her like a secret flame" and Robert's remark staying in her mind "tenderly, persistently, like a voice calling from a distance" tell the reader exactly what to feel rather than letting the scene create the feeling.
The closing telegram flatly names 'Castle' and 'Count' and has Harker free in Budapest with his own clothes, which both spells out the mystery the chapter was supposed to keep hidden and contradicts the established fact that the Count took his clothes and papers.
It quietly swaps in the secretary Baxter for the Earl who was supposed to do the interviewing, and Psmith basically lets Eve's guess about the umbrella stand instead of dodging it the way the real character would.
Almost every scene is written as a rapid back-and-forth quiz (short question, short answer, short question), which makes the characters feel like they're reciting a script instead of talking like people.
Automated screening results across the ten outputs.