The Standings → GPT‑6 Astra (low)
OpenAI
low reasoning effort · via Codex CLI · judged by Claude Sonnet 5 · not the setting shown on the standings; also run at high73.0 medium70.1 xhigh68.6
Results from run ‘first’: ten items, ten verdicts.
GPT‑6 Astra (low) scores 71 on craft, 8 points above the field’s 63 and 22 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 · 2,316 words on average · est. generation cost $3.59
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 71. 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 Renfield's scene is sharply drawn, showing his calculating, self-interested logic through his own carefully composed job application rather than through explanation.
Worst The invented mystery of an impersonator stealing Jonathan's bag feels like a bolted-on detective-story twist rather than something growing naturally out of the Gothic story already begun.
Best The plotting is tidy and satisfying — the umbrella's journey from Eve back to Walderwick to Psmith makes logical sense, and the gardening-catalogue mix-up in the interview is a neat, in-character piece of business.
Worst The dialogue is clipped and efficient where Wodehouse would spin a joke out at length, so the humour lands but rarely builds to a big comic payoff.
Best It captures Anne's marginal status in the family with real precision, especially the line about her father turning to her 'with the air of a man to whom an observation has been addressed by mistake.'
Worst The scene is a touch flatter and more subdued than it needs to be, so the comic and emotional peaks land less forcefully than they could.
Best The widow's interview is tightly controlled and believably evasive, letting her character come through in what she dodges rather than in what she announces about herself.
Worst The interrogation scenes lean so heavily on clipped one-line question-and-answer volleys that they start to feel like a transcript rather than a scene with any texture or surprise.
Best It picks up in the very instant the source chapter leaves off, mid-walk with the wife, which makes the opening feel seamless.
Worst The sentences and dialogue are plainer and more modern-sounding than Wells's actual style, so it reads less like a period narrator and more like generic literary prose.
Best The return to Georgette is genuinely warm — her laughing for once without hiding her teeth gives the chapter its most human beat.
Worst Many of the back-and-forth conversations sound alike, so the middle of the chapter drifts instead of pulling the reader forward.
Best It is admirably restrained and controlled, especially in Betteredge's refusal to pry into the packet and the quiet handling of Lady Verinder's prohibition.
Worst It rushes through the required events so efficiently that the chapter reads like a checklist being ticked off rather than a scene unfolding at its own pace.
Best It stays admirably restrained and lets small everyday details, like the fly resettling on the same word or the dropped key, carry the emotional weight instead of spelling it out.
Worst Several required moments (the dinner conversation about the telegram, the fishing plan) are reported in summary rather than played out as scenes, so the chapter feels thinner and less alive than it should.
Best Poirot's physical examination of the room is precise and tactile, with small convincing procedural details like measuring, kneeling to look under furniture, and using tweezers.
Worst The chapter opens with a long question-and-answer recap of events the reader already witnessed in the previous chapter, which stalls momentum before anything new happens.
Best It stays modest and unpretentious, and small human touches, like the servant Diggory's cheerful, patient Yorkshire greeting, land well.
Worst The final lines flatly state that Mary decided to find out what the sound was, which reads like a checklist being ticked off rather than a moment she arrives at herself.
Headline, compliance and craft for each chapter, 0–100.
| Item | Headline | Compliance | Craft | Words |
|---|---|---|---|---|
| Dracula item 015 · after ch. 5 · Bram Stoker | 76.9 | 100 | 77 | 2,265 |
| Leave It to Psmith item 022 · after ch. 3 · P. G. Wodehouse | 76.9 | 100 | 77 | 2,568 |
| Persuasion item 001 · after ch. 1 · Jane Austen | 73.9 | 95 | 78 | 2,527 |
| Red Harvest item 023 · after ch. 2 · Dashiell Hammett | 73.2 | 100 | 73 | 2,294 |
| The War of the Worlds item 006 · after ch. 1 · H. G. Wells | 68.9 | 95 | 72 | 2,439 |
| The Sun Also Rises item 021 · after ch. 3 · Ernest Hemingway | 68.8 | 100 | 69 | 2,253 |
| The Moonstone item 011 · after ch. 8 · Wilkie Collins | 67.5 | 100 | 68 | 2,485 |
| The Awakening item 016 · after ch. 7 · Kate Chopin | 66.5 | 100 | 66 | 1,625 |
| The Mysterious Affair at Styles item 017 · after ch. 3 · Agatha Christie | 65.9 | 95 | 69 | 2,473 |
| The Secret Garden item 013 · after ch. 2 · Frances Hodgson Burnett | 63.4 | 100 | 63 | 2,232 |
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’.
Several required moments (the dinner conversation about the telegram, the fishing plan) are reported in summary rather than played out as scenes, so the chapter feels thinner and less alive than it should.
The invented mystery of an impersonator stealing Jonathan's bag feels like a bolted-on detective-story twist rather than something growing naturally out of the Gothic story already begun.
The interrogation scenes lean so heavily on clipped one-line question-and-answer volleys that they start to feel like a transcript rather than a scene with any texture or surprise.
The chapter opens with a long question-and-answer recap of events the reader already witnessed in the previous chapter, which stalls momentum before anything new happens.
Automated screening results across the ten outputs.