NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYCR5Z9PK44RPSNZEB8GZ3A4
Run Details
Model:
Claude Opus 5 (Reasoning)
Test:
Bad Writing Habits
Scenario:
Mystery: examining a crime scene
Duration:
1m 26s
Cost:
$0.122900
Provider:
Amazon Bedrock
Evaluation summary
Score
Evaluator
Details Link
100.00%
Adverbs in dialogue tags
(details)
88.28%
AI-ism adverb frequency
(details)
100.00%
AI-ism character names
(details)
100.00%
AI-ism location names
(details)
94.14%
AI-ism word frequency
(details)
100.00%
Cliché density
(details)
100.00%
Emotion telling (show vs. tell)
(details)
100.00%
Filter word density
(details)
100.00%
Gibberish response detection
(details)
100.00%
Markdown formatting overuse
(details)
100.00%
Missing dialogue indicators (quotation marks)
(details)
50.00%
Name drop frequency
(details)
100.00%
Narrator intent-glossing
(details)
100.00%
"Not X but Y" pattern overuse
(details)
100.00%
Overuse of "that" (subordinate clause padding)
(details)
100.00%
Paragraph length variance
(details)
93.06%
Passive voice overuse
(details)
100.00%
Past progressive (was/were + -ing) overuse
(details)
4.93%
Em-dash & semicolon overuse
(details)
100.00%
Purple prose (modifier overload)
(details)
100.00%
Repeated phrase echo
(details)
100.00%
Sentence length variance
(details)
74.47%
Sentence opener variety
(details)
100.00%
Adverb-first sentence starts
(details)
100.00%
Pronoun-first sentence starts
(details)
100.00%
Subject-first sentence starts
(details)
0.00%
Subordinate conjunction sentence starts
(details)
100.00%
Technical jargon density
(details)
100.00%
Useless dialogue additions
(details)
100.00%
Dialogue tag variety (said vs. fancy)
(details)
90.1626%
Overall
Model input
System
You are an expert novelist with a gift for vivid, immersive prose. You write in tight third-person limited point of view with a strong narrative voice. Your prose is clean, precise, and engaging — you favor strong verbs, concrete details, and natural-sounding dialogue. You will be given a story bible (codex) containing canonical details about characters, locations, items, and lore. Use these details to ground your writing in the established world. Stay consistent with the codex. Write only the prose of the scene. Do not include titles, chapter headings, markdown formatting, or any commentary before or after the scene.
User/Human
<codex> <character id="4416" name="Detective Harlow Quinn" aliases="Quinn, Detective Quinn, Harlow" pronouns="she/her" species="Human" storyRole="Antagonist"> <eyeColor>brown</eyeColor> <hair>closely cropped salt-and-pepper</hair> <height>5'9"</height> <age>41</age> <jawline>sharp jaw</jawline> <watch>worn leather watch on her left wrist</watch> <bearing>military precision</bearing> <description>A Metropolitan Police detective with 18 years of decorated service who suspects the clique is involved in criminal activity.</description> <backstory>She lost her partner, DS Morris, during a case three years ago under unexplained circumstances that had supernatural origins she doesn't yet understand.</backstory> </character> <character id="4414" name="Eva Kowalski" aliases="Eva, Evie" pronouns="she/her" species="Human" storyRole="Best Friend"> <eyeColor>green</eyeColor> <hair>curly red</hair> <height>5'4"</height> <age>26</age> <glasses>round glasses</glasses> <satchel>worn leather satchel full of books</satchel> <freckles>freckled complexion</freckles> <nervousHabit>tucking hair behind her left ear</nervousHabit> <description>Aurora's childhood friend and an occult researcher.</description> <backstory>Moved to London two years before Aurora and works as a research assistant at the British Museum's restricted archives. She holds a master's degree in Ancient History from Oxford.</backstory> </character> <location id="856" name="The Veil Market" aliases="Veil Market, the Market, underground market" realm="Earth"> <currentLocation>abandoned Tube station beneath Camden</currentLocation> <entryRequirement>bone token</entryRequirement> <schedule>moves locations every full moon</schedule> <description>A hidden supernatural black market that sells enchanted goods, banned alchemical substances, and information.</description> </location> <item id="2004" name="Veil Compass" aliases="the Compass, shadow compass" type="Tool" origin="The Veil Market"> <casingMaterial>small brass compass</casingMaterial> <faceMarkings>face etched with protective sigils</faceMarkings> <needle>points toward the nearest supernatural rift or portal</needle> <patina>casing has a patina of verdigris</patina> <creator>crafted by a Shade artisan</creator> <description>A compass attuned to supernatural energy.</description> </item> </codex> Write a fully fleshed out scene of at least 1000 words in the mystery genre. A detective arrives at a crime scene in an unusual location. Something is wrong — the evidence doesn't add up. A colleague is present, offering a different interpretation. Through careful observation and deduction, the detective begins to see what others have missed. Use the codex entries above as your world reference for characters and setting. Only include the prose of the scene without any markdown or commentary.
Model response
Messages
Meta
AI
The hoarding at the bottom
of the service stair had
been screwed
shut from the inside, and that was the first thing that was wrong.
Detective Harlow Quinn crouched in
front of it with her torch held low and raking, the way you held a light when you wanted the truth out of a surface. Eight screws. Modern Phillips heads, galvanised, bright as new pins in the beam.
The plywood around them was
forty years soft, furred with damp, and yet the screw threads had bitten clean.
Somebody had put them in
within the last month.
Somebody standing on this side
of the barrier, in a stairwell that dead-ended at a bricked arch and a hundred feet of London clay. "Kids," said DC Ashworth from three steps up. "They've got a way in we haven't found. There's always a way in." "Then find it." "That's what the search team's for." Quinn straightened.
Her knees made a sound
she chose to ignore. Forty-one years, eighteen of them on the job, and she'd
learned that the
noises a body made were only evidence if you were the one lying down.
The platform beyond was Northern
Line, decommissioned in 1924 and never quite forgotten: oxblood tiles, cream tiles, the old station name half-scoured off the wall so it read only ND TOWN. Cabling looped the ceiling like veins.
Dust lay everywhere — the
particular dust of the deep tunnels, a grey-black flour of brake iron and skin cells and coal smoke ninety years cold. Everywhere except the middle.
Quinn stood at the platform's
edge and let her torch travel down its length. Down the centre, running the full eighty yards from the tunnel mouth to the blocked-off lift shaft, the dust had
been worn
away. Not scuffed. Not tracked. Polished.
A broad ribbon of bare
tile with a shine to it, the way the flagstones wore in the nave of a church, the way a doorstep wore under a century of boots. "Hundreds," she said. "Sorry, ma'am?" "That's a hundred years of
footfall
or it's hundreds of pairs of feet. And this station's been sealed since before my grandfather was born."
She stepped down onto the
polished strip.
Her own shoe left nothing
at all. "So it's hundreds of feet, and recently, and repeatedly." "Rave,"
Ashworth said, warming to it
. "Illegal party. You get them in the old stations, the Mail Rail, the reservoirs. Sound system, glowsticks, four hundred quid a door." "Show me one can." He didn't answer. "One bottle," Quinn said. "One cigarette end. One condom, one crisp packet, one smear of vomit. Four hundred people in a sealed hole in the ground and not so much as a chewing-gum wrapper."
She turned in a slow
circle, torch level. "I've done raves, Ashworth. They leave a tidemark you can read like a river." What this place had instead were holes.
She found the first without
meaning to, her heel finding air.
A neat circular bore in
the tile, an inch across, tapped clean. Then another, four feet on.
Then two more, squared off
with the first pair.
She walked the ribbon and
counted them, and the pattern came up out of the floor at her like handwriting: pitches. Stall pitches, four uprights each, laid out in a double row with a walkway between. Beside the bores, chalked onto the tile in a small tidy hand, numbers. Not sequential. 19. 4. 87. 12.
Someone's system, and not a
system meant for her. Wax, too. Fat runnels of it welded to the tiles, white and green and a dark red that took a scrape of her thumbnail to prove wasn't blood. Beeswax. She smelled her thumb: honey, and under it, cardamom. Burnt sugar. Something animal and warm and gone, like a stable a week after the horses. "Ma'am." Ashworth had his notebook out and hadn't written anything in it. "The body." The body lay at the tunnel mouth, half in and half out of the dark, and the pathologist had come and gone and left the poor man tented in blue. Quinn lifted the sheet. Male, fifties, wiry, dressed in a wool coat too heavy for June.
His face was turned toward
the tunnel.
His right hand was open
and empty; his left
was clenched
so hard the nails had broken the skin of his palm. There was a wound below his sternum, a wound she'd normally call a wound — a puncture, narrow, deep, and clean. There was no blood. Not on the coat. Not on the tile. Not a drop, not a spatter, not so much as the rusty crust a wound weeps once the heart's given up. The cloth around the tear was dry as a page. "He was killed elsewhere and dumped," Ashworth said. "Dumped, cleaned, whatever. Drugs, most likely. There's a supply line runs through Camden, we've had three—" "He walked in." "You can't know—" "He walked in and he didn't walk out." Quinn tracked the torch beam along the polished ribbon toward the tunnel. There, at the
very
lip where clean tile gave way to ballast and grease, the dust resumed — thin, but there. And in it: one set of prints. Brogues, worn on the outside edge, size nine or ten. Coming out of the tunnel. Toward the platform. None going back. "So he came from down the line." "Everyone did."
She swept the beam wide
. In the ballast beyond the tunnel mouth, chaos: gravel
churned
, dragged, kicked into ridges. A herd's worth. And all of it moving away, into the black, southbound. "The whole lot of them went that way. Fast. He was the one who stopped." Behind her a woman's voice said, "You'll want to be careful past the tunnel mouth." Quinn turned. The uniform on the stair
was already apologising
, and beside him stood a young woman in a wax jacket much too big for her, five foot four in flat boots, red hair going every direction the damp could take it, round glasses fogged at the rims.
She had a leather satchel
on her hip, stuffed to bursting with books, and she
was holding
it like a shield. "Eva Kowalski. British Museum. I'm — the Department asked for someone on the markings, I do restricted archives, palaeography mostly—" She stopped. Tucked a curl behind her left ear. "I'm sorry. There's a body." "There is." Quinn let the sheet down
gently
. "You're the epigrapher." "I'm a research assistant." "They sent you at four in the morning." "I answered the phone." Eva's eyes went past her to the wall by the lift shaft, and Quinn watched them go, and watched them stop, and watched something in the girl's face let go of a breath it had been holding. Quinn had been
saving that wall
.
It was covered — the
whole tiled pier of it, floor to eight feet — in figures scratched and chalked and, in one place, burnt. Circles inside circles. Barbed spirals. A thing like an eye that wasn't an eye. "Well," Quinn said. "Tell me what I'm looking at." Eva came closer, careful of the polished strip, and stood with her hands behind her back like a child at a museum. "It's — apotropaic. Protective. This one's a variant of a Coptic ward, you find it on lintels in Upper Egypt. This cluster's much later, seventeenth century, English cunning-folk, they'd chalk it under the eaves against—"
She caught herself
. "The point is, none of it's a summons. It's the opposite. Whoever put this here was keeping something out." "Or keeping something in." "That's not how the grammar works." "You read it
very
fast," Quinn said. Eva's chin came up half an inch. "I've seen the type before." "You read it before you looked at it." Quinn kept her voice mild, pleasant, the voice she used on witnesses who were about to tell her something true by accident. "You came down those stairs and your eyes went to that wall before I said a word about markings. You knew what was on it. You knew where it was." "I—" The hand went to the left ear again. "It's a known site. In the literature." "Which literature? I'll get a warrant for it." Nothing. Behind the fogged glasses, green eyes, and the particular stillness of a clever person deciding how much to spend. Quinn crouched again — knees, again — beside the dead man's left hand, and with two gloved fingers she opened it. A disc of bone, yellowed, drilled through, the size of a ten-pence piece. Snapped clean in half.
Only half was in his
palm. And under the bench behind him, where it had rolled and lodged against a rivet, something caught her torch and threw it back green. She fished it out. A compass, brass, small enough to hide in a fist, the casing gone soft and crusted with verdigris. Sigils cut into the face — she looked up at the wall, and there they were again, two of them, the barbed spiral and the eye that wasn't.
She turned it flat on
her palm. The needle swung north, hesitated, and then walked steadily round to point south, down the tunnel, into the dark, and held there like a dog on a scent. Quinn turned a slow half-circle. The needle held south. Behind her, Eva Kowalski said,
very
quietly, "Please don't follow that." Quinn glanced at her left wrist, at the worn leather
watch
she'd worn for eleven years, since before Morris, since before the night she'd found his car door open and his radio still warm and nothing else, nothing else at all. The
watch
had stopped at 2:14. Full moon had risen at 2:11. "Ashworth," Quinn said, closing her hand around the compass. "Get me the lunar calendar for the last three years. Every full moon. And every unexplained body found within half a mile of a disused station on the morning after." "Ma'am, that's—" "That's the pattern." She stood. The needle pressed against her palm, insistent, southbound. "They pack up and they move, and once in a while somebody's too slow. He wasn't dumped here, Ashworth. He was left behind."
She looked at Eva, who
had gone the colour of the tiles. "Now," Quinn said. "Which of us wants to tell the truth first?"
created_at
"2026-07-25T13:42:06.398Z"
7/25/2026, 1:42:06 PM
model
"anthropic/claude-opus-5-20260723"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
4317
moderation_latency
(null)
generation_time
85601
tokens_prompt
896
tokens_completion
3106
native_tokens_prompt
1285
native_tokens_completion
4659
native_tokens_completion_images
(null)
native_tokens_reasoning
676
native_tokens_cached
0
num_media_prompt
(null)
num_input_audio_prompt
(null)
num_media_completion
0
num_search_results
(null)
num_fetches
(null)
web_search_engine
(null)
origin
"https://ncbench.com/"
is_byok
false
finish_reason
"stop"
native_finish_reason
"end_turn"
service_tier
"default"
usage
0.1229
router
(null)
provider_responses
0
endpoint_id
"76cb4608-f48c-483d-8da8-9957fb44244e"
id
"msg_011CdNsZwknn9Md7u9KjHk1J"
is_byok
false
latency
1791
model_permaslug
"anthropic/claude-opus-5-20260723"
provider_name
"Amazon Bedrock"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.18.0; linux; x64))"
http_referer
(null)
request_id
"req-1784986926-Quyxe0uxH4E7miD8JWSH"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1784986926-6Q3CaHes8s5g1XEFP9XJ"
upstream_id
"msg_011CdNsZwknn9Md7u9KjHk1J"
total_cost
0.1229
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Amazon Bedrock"
response_cache_source_id
(null)
data_region
"global"
Evaluation details
Result
Evaluator
Details
Meta Data
100.00%
Adverbs in dialogue tags
Target: ≤10% dialogue tags with adverbs
totalTags
23
adverbTagCount
1
adverbTags
0
"Behind her Eva Kowalski said very [very]"
dialogueSentences
53
tagDensity
0.434
leniency
0.868
rawRatio
0.043
effectiveRatio
0.038
88.28%
AI-ism adverb frequency
Target: <2% AI-ism adverbs (58 tracked)
wordCount
1707
totalAiIsmAdverbs
4
found
0
adverb
"very"
count
3
1
adverb
"gently"
count
1
highlights
0
"very"
1
"gently"
100.00%
AI-ism character names
Target: 0 AI-default names (17 tracked, −20% each)
codexExemptions
(empty)
found
(empty)
100.00%
AI-ism location names
Target: 0 AI-default location names (33 tracked, −20% each)
codexExemptions
(empty)
found
(empty)
94.14%
AI-ism word frequency
Target: <2% AI-ism words (290 tracked)
wordCount
1707
totalAiIsms
2
found
0
word
"footfall"
count
1
1
word
"churned"
count
1
highlights
0
"footfall"
1
"churned"
100.00%
Cliché density
Target: ≤1 cliche(s) per 800-word window
totalCliches
0
maxInWindow
0
found
(empty)
highlights
(empty)
100.00%
Emotion telling (show vs. tell)
Target: ≤3% sentences with emotion telling
emotionTells
0
narrationSentences
115
matches
(empty)
100.00%
Filter word density
Target: ≤3% sentences with filter/hedge words
filterCount
1
hedgeCount
0
narrationSentences
115
filterMatches
0
"watch"
hedgeMatches
(empty)
100.00%
Gibberish response detection
Target: ≤1% gibberish-like sentences (hard fail if a sentence exceeds 800 words)
analyzedSentences
145
gibberishSentences
0
adjustedGibberishSentences
0
longSentenceCount
0
runOnParagraphCount
0
giantParagraphCount
0
wordSaladCount
0
repetitionLoopCount
0
controlTokenCount
0
repeatedSegmentCount
0
maxSentenceWordsSeen
46
ratio
0
matches
(empty)
100.00%
Markdown formatting overuse
Target: ≤5% words in markdown formatting
markdownSpans
0
markdownWords
0
totalWords
1717
ratio
0
matches
(empty)
100.00%
Missing dialogue indicators (quotation marks)
Target: ≤10% speech attributions without quotation marks
totalAttributions
21
unquotedAttributions
0
matches
(empty)
50.00%
Name drop frequency
Target: ≤1.0 per-name mentions per 100 words
totalMentions
36
wordCount
1259
uniqueNames
12
maxNameDensity
1.43
worstName
"Quinn"
maxWindowNameDensity
3.5
worstWindowName
"Quinn"
discoveredNames
Harlow
1
Quinn
18
Phillips
1
London
1
Ashworth
4
Northern
1
Line
1
June
1
Tucked
1
Eva
5
Kowalski
1
Morris
1
persons
0
"Harlow"
1
"Quinn"
2
"Phillips"
3
"Ashworth"
4
"Line"
5
"Eva"
6
"Kowalski"
7
"Morris"
places
0
"London"
globalScore
0.785
windowScore
0.5
100.00%
Narrator intent-glossing
Target: ≤2% narration sentences with intent-glossing patterns
analyzedSentences
72
glossingSentenceCount
0
matches
(empty)
100.00%
"Not X but Y" pattern overuse
Target: ≤1 "not X but Y" per 1000 words
totalMatches
0
per1kWords
0
wordCount
1717
matches
(empty)
100.00%
Overuse of "that" (subordinate clause padding)
Target: ≤2% sentences with "that" clauses
thatCount
2
totalSentences
145
matches
0
"learned that the"
1
"saving that wall"
100.00%
Paragraph length variance
Target: CV ≥0.5 for paragraph word counts
totalParagraphs
65
mean
26.42
std
24.94
cv
0.944
sampleLengths
0
25
1
95
2
21
3
3
4
6
5
38
6
63
7
4
8
74
9
3
10
2
11
47
12
28
13
4
14
3
15
56
16
7
17
101
18
54
19
14
20
30
21
4
22
62
23
4
24
39
25
24
26
3
27
3
28
69
29
7
30
49
31
15
32
69
33
34
34
11
35
4
36
8
37
41
38
6
39
39
40
9
41
79
42
4
43
6
44
7
45
12
46
60
47
16
48
8
49
20
93.06%
Passive voice overuse
Target: ≤2% passive sentences
passiveCount
4
totalSentences
115
matches
0
"been screwed"
1
"been worn"
2
"was clenched"
3
"was covered "
100.00%
Past progressive (was/were + -ing) overuse
Target: ≤2% past progressive verbs
pastProgressiveCount
2
totalVerbs
197
matches
0
"was already apologising"
1
"was holding"
4.93%
Em-dash & semicolon overuse
Target: ≤2% sentences with em-dashes/semicolons
emDashCount
8
semicolonCount
1
flaggedSentences
7
totalSentences
145
ratio
0.048
matches
0
"Dust lay everywhere — the particular dust of the deep tunnels, a grey-black flour of brake iron and skin cells and coal smoke ninety years cold."
1
"His right hand was open and empty; his left was clenched so hard the nails had broken the skin of his palm."
2
"There was a wound below his sternum, a wound she'd normally call a wound — a puncture, narrow, deep, and clean."
3
"There, at the very lip where clean tile gave way to ballast and grease, the dust resumed — thin, but there."
4
"It was covered — the whole tiled pier of it, floor to eight feet — in figures scratched and chalked and, in one place, burnt."
5
"Quinn crouched again — knees, again — beside the dead man's left hand, and with two gloved fingers she opened it."
6
"Sigils cut into the face — she looked up at the wall, and there they were again, two of them, the barbed spiral and the eye that wasn't."
100.00%
Purple prose (modifier overload)
Target: <4% adverbs, <2% -ly adverbs, no adj stacking
wordCount
1258
adjectiveStacks
0
stackExamples
(empty)
adverbCount
32
adverbRatio
0.025437201907790145
lyAdverbCount
7
lyAdverbRatio
0.005564387917329093
100.00%
Repeated phrase echo
Target: ≤20% sentences with echoes (window: 2)
totalSentences
145
echoCount
0
echoWords
(empty)
100.00%
Sentence length variance
Target: CV ≥0.4 for sentence word counts
totalSentences
145
mean
11.84
std
10.17
cv
0.859
sampleLengths
0
25
1
30
2
2
3
11
4
19
5
9
6
24
7
8
8
13
9
3
10
6
11
2
12
9
13
27
14
31
15
6
16
26
17
4
18
14
19
23
20
2
21
2
22
1
23
32
24
3
25
2
26
31
27
7
28
9
29
6
30
22
31
4
32
3
33
4
34
38
35
14
36
7
37
11
38
12
39
5
40
9
41
21
42
15
43
13
44
2
45
1
46
1
47
1
48
1
49
9
74.47%
Sentence opener variety
Target: ≥60% unique sentence openers
consecutiveRepeats
10
diversityRatio
0.48936170212765956
totalSentences
141
uniqueOpeners
69
100.00%
Adverb-first sentence starts
Target: ≥3% sentences starting with an adverb
adverbCount
4
totalSentences
91
matches
0
"Everywhere except the middle."
1
"Then another, four feet on."
2
"Then two more, squared off"
3
"Only half was in his"
ratio
0.044
100.00%
Pronoun-first sentence starts
Target: ≤30% sentences starting with a pronoun
pronounCount
17
totalSentences
91
matches
0
"Her knees made a sound"
1
"She stepped down onto the"
2
"Her own shoe left nothing"
3
"He didn't answer."
4
"She turned in a slow"
5
"She found the first without"
6
"She walked the ribbon and"
7
"She smelled her thumb: honey,"
8
"His face was turned toward"
9
"His right hand was open"
10
"She swept the beam wide"
11
"She had a leather satchel"
12
"It was covered — the"
13
"She caught herself"
14
"She fished it out."
15
"She turned it flat on"
16
"She looked at Eva, who"
ratio
0.187
100.00%
Subject-first sentence starts
Target: ≤72% sentences starting with a subject
subjectCount
60
totalSentences
91
matches
0
"The hoarding at the bottom"
1
"Detective Harlow Quinn crouched in"
2
"The plywood around them was"
3
"Somebody had put them in"
4
"Somebody standing on this side"
5
"Her knees made a sound"
6
"The platform beyond was Northern"
7
"Dust lay everywhere — the"
8
"Quinn stood at the platform's"
9
"A broad ribbon of bare"
10
"She stepped down onto the"
11
"Her own shoe left nothing"
12
"Ashworth said, warming to it"
13
"He didn't answer."
14
"She turned in a slow"
15
"She found the first without"
16
"A neat circular bore in"
17
"She walked the ribbon and"
18
"Stall pitches, four uprights each,"
19
"Someone's system, and not a"
ratio
0.659
0.00%
Subordinate conjunction sentence starts
Target: ≥2% sentences starting with a subordinating conjunction
subConjCount
0
totalSentences
91
matches
(empty)
ratio
0
100.00%
Technical jargon density
Target: ≤6% sentences with technical-jargon patterns
analyzedSentences
41
technicalSentenceCount
0
matches
(empty)
100.00%
Useless dialogue additions
Target: ≤5% dialogue tags with trailing filler fragments
totalTags
23
uselessAdditionCount
0
matches
(empty)
100.00%
Dialogue tag variety (said vs. fancy)
Target: ≤10% fancy dialogue tags
totalTags
10
fancyCount
0
fancyTags
(empty)
dialogueSentences
53
tagDensity
0.189
leniency
0.377
rawRatio
0
effectiveRatio
0
90.1626%