NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-17-01KHMR25JMEWXZ9SV7F73M05AG
Run Details
Model:
Claude Sonnet 4.5
Test:
Bad Writing Habits
Scenario:
Mystery: examining a crime scene
Duration:
31.7s
Cost:
$0.026226
Provider:
Amazon Bedrock
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 abandoned Camden Tube station stank of rust and stagnant water. Detective Harlow Quinn descended the last metal staircase, her boots
echoing
against concrete that hadn't seen maintenance since the station closed in 1952. The beam from her torch cut through darkness thick as wool, catching on Victorian-era tile work and decades of grime. She'd received the anonymous tip three hours ago. Body in the old station. Come alone. Quinn never came alone. "Christ, it's like a tomb down here." DS
Patel
's voice bounced off the curved walls behind her. Young, eager, still believed the job was about justice rather than damage control. "Watch your step," Quinn said. "The platforms are crumbling." They emerged onto what had once been the northbound platform. Someone had set up battery-powered work lights in a rough semicircle, casting harsh shadows across the scene. The body lay near the platform edge, a man in his thirties wearing expensive trainers and a leather jacket that had seen better days. His wallet sat open beside him, cards and cash untouched. Quinn knelt, careful not to disturb the perimeter. She'd learned military precision during her years in the Royal Military Police, and eighteen years with the Met hadn't dulled those instincts. The victim's eyes stared at the vaulted ceiling, clouded but not yet fully glazed. Recent, then. Maybe six hours. "No visible wounds,"
Patel
said, circling to the other side. "Heart attack?" "At thirty-five?" Quinn shifted her torch beam across the body. "Look at his hands." The fingers
were splayed
, knuckles scraped raw. Defensive wounds, or something else. The man's mouth hung open in a silent scream, and when Quinn leaned closer, she caught the faint smell beneath the station's must — something like ozone and copper, but wrong somehow. Twisted. Her left hand moved unconsciously to the worn leather
watch
on her wrist. Three years since Morris died, and she still checked the time by touch when something felt off. "Detective Quinn?"
Patel
held up an evidence bag. "Found this under the body." Quinn took the bag and held it to the light. A small brass compass, its face
etched
with symbols she didn't recognize. Protective sigils, maybe, if she believed in that sort of thing. The casing had a green patina of verdigris, and when she tilted the bag, the needle swung — not toward
magnetic
north, but toward the tunnel mouth behind her. "Bag it," she said. "And get photos of e
very
thing. The positioning, the scrapes on his hands, that discoloration on the platform." She stood and swept her torch across the wider scene. The work lights only illuminated about twenty feet in each direction, but beyond their reach, the platform str
etched
into shadow. And there — her beam caught something on the far wall. Scratches in the grime, deliberate marks forming patterns. Quinn walked toward them, her
footsteps
hollow. Behind her,
Patel
's camera flash strobed across the body. The marks weren't graffiti. Too precise, too intentional. They formed a rough circle with radiating lines, like a child's drawing of the sun. Or a diagram. And at the center, that same symbol she'd seen on the compass. "
Patel
." She kept her voice level. "When did the tip come in?" "Logged at oh-four-hundred hours." "And we're the first responders?" "Yes, ma'am. Dispatch said you requested—" "I didn't request anything." Silence, punctuated by water dripping somewhere in the tunnels. Quinn's jaw tightened. Someone had called this in specifically for her. Someone who knew she'd come, who knew she'd see what others might miss. She turned back to the body, studying the scene with fresh eyes. The wallet open but nothing stolen. The expensive trainers caked with tunnel mud, suggesting he'd walked down here rather than
being dumped
. The compass tucked beneath him as if he'd clutched it to his chest. And those scrapes on his hands — Quinn crouched again, this time examining the platform edge near the body. Fresh gouges in the concrete, fingernail width. She played her torch along the edge toward the tunnel mouth. More gouges, a trail of them, as if someone had
tried to
pull themselves away from something. "He was dragged," she said. "But there's no—"
Patel
stopped. "There's no drag marks on the body. No scuffing on the jacket." "No." Quinn stood, following the gouge marks with her torch. They led to the platform edge, then stopped at a section of wall that looked darker than the rest. Water damage, maybe. Or — She pressed her palm against it. Cold, far colder than the surrounding stone. And that smell again, ozone and copper and something underneath that made her teeth ache. "Ma'am, we should wait for the full forensics team." Quinn pulled her hand back. Her palm tingled, pins and needles spreading up her wrist. She'd felt something like this before, three years ago, the night Morris died. They'd been investigating a series of break-ins at historic sites, following a lead to an old church in Southwark. Morris had gone ahead, checking the crypt. When Quinn found him ten minutes later, he was already cold, eyes wide with the same expression this victim wore. Heart attack, the coroner said. Thirty-eight years old, fit as an ox, dead from heart failure. The compass needle had swung back toward the body. Not quite pointing at it, but toward the space where it lay. Or through it, toward something else. Quinn turned to
Patel
. "I want this entire station
processed
. E
very
inch, both platforms, the tunnels within a hundred meters. And I want to know who owns this compass." "What are we looking for?" Good question. Evidence of what? She thought of Morris, of his cold hand in hers, of the way the case had
been closed
so
quickly
. Thought of the symbols on the wall, the impossible chill of that stone, the way her instincts
screamed that something
was
very
wrong with this scene. "Anomalies," she said finally. "Anything that doesn't fit the pattern."
Patel
nodded and returned to photographing the body. Quinn walked back to the wall marks, pulling out her phone to document them. The camera flash washed out the symbols, turning them into meaningless scratches. But in the torch light, they almost
seemed to
pulse. Her
watch
felt heavy on her wrist. She checked it without thinking: half past seven. Sunrise soon, not that it mattered down here in the dark. The anonymous tipster had wanted her to see this. Had known she'd understand, or at least suspect. Someone
was leaving
her a trail of breadcrumbs, and she was finally beginning to see where they might lead.
created_at
"2026-02-17T02:50:25.698828+00:00"
2/17/2026, 2:50:25 AM
model
"anthropic/claude-4.5-sonnet-20250929"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2814
moderation_latency
205
generation_time
31457
tokens_prompt
923
tokens_completion
1706
native_tokens_prompt
937
native_tokens_completion
1561
native_tokens_completion_images
(null)
native_tokens_reasoning
0
native_tokens_cached
0
num_media_prompt
(null)
num_input_audio_prompt
(null)
num_media_completion
0
num_search_results
(null)
origin
"https://ncbench.com/"
is_byok
false
finish_reason
"stop"
native_finish_reason
"stop"
usage
0.026226
router
(null)
provider_responses
0
id
"176613a2-03db-4809-8ce8-76ed2f89355c"
status
200
is_byok
false
latency
2814
endpoint_id
"d8bee848-f020-454a-b2be-4a7ce868f99f"
provider_name
"Amazon Bedrock"
model_permaslug
"anthropic/claude-4.5-sonnet-20250929"
api_type
"completions"
id
"gen-1771296593-Fvr2WJOMzrfTTnAFMitK"
upstream_id
"176613a2-03db-4809-8ce8-76ed2f89355c"
total_cost
0.026226
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Amazon Bedrock"
Evaluation details
Result
Evaluator
Details
Meta Data
66.6667%
Adverbs in dialogue tags
Target: ≤10% dialogue tags with adverbs
totalTags
5
adverbTagCount
1
adverbTags
0
"she said finally [finally]"
81.8182%
AI-ism adverb frequency
Target: <1% AI-ism adverbs (57 tracked)
wordCount
1100
totalAiIsmAdverbs
2
found
0
adverb
"quickly"
count
1
1
adverb
"very"
count
1
highlights
0
"quickly"
1
"very"
80.0000%
AI-ism character names
Target: 0 AI-default names (17 tracked, −20% each)
codexExemptions
(empty)
found
0
"Patel"
100.0000%
AI-ism location names
Target: 0 AI-default location names (33 tracked, −20% each)
codexExemptions
(empty)
found
(empty)
54.5455%
AI-ism word frequency
Target: <1% AI-ism words (140 tracked)
wordCount
1100
totalAiIsms
5
found
0
word
"processed"
count
1
1
word
"magnetic"
count
1
2
word
"footsteps"
count
1
3
word
"echoing"
count
1
4
word
"etched"
count
1
highlights
0
"processed"
1
"magnetic"
2
"footsteps"
3
"echoing"
4
"etched"
100.0000%
Cliche density
Target: ≤1 cliche(s) per 800-word window
totalCliches
0
maxInWindow
0
found
(empty)
highlights
(empty)
100.0000%
Emotion telling (show vs. tell)
Target: ≤3% sentences with emotion telling
emotionTells
0
narrationSentences
94
matches
(empty)
100.0000%
Filter word density
Target: ≤12% sentences with filter/hedge words
filterCount
2
hedgeCount
2
narrationSentences
94
filterMatches
0
"watch"
hedgeMatches
0
"tried to"
1
"seemed to"
100.0000%
Overuse of "that" (subordinate clause padding)
Target: ≤10% sentences with "that" clauses
thatCount
1
totalSentences
109
matches
0
"screamed that something"
100.0000%
Paragraph length variance
Target: CV ≥0.5 for paragraph word counts
totalParagraphs
40
mean
27.45
std
19.15
cv
0.697
sampleLengths
0
54
1
15
2
4
3
30
4
9
5
61
6
49
7
12
8
14
9
45
10
30
11
13
12
62
13
21
14
49
15
16
16
38
17
12
18
4
19
5
20
6
21
4
22
33
23
54
24
47
25
5
26
17
27
34
28
28
29
9
30
74
31
16
32
27
33
29
34
5
35
51
36
10
37
44
38
26
39
36
100.0000%
Passive voice overuse
Target: ≤5% passive sentences
passiveCount
3
totalSentences
94
matches
0
"were splayed"
1
"being dumped"
2
"been closed"
100.0000%
Past progressive (was/were + -ing) overuse
Target: ≤10% past progressive verbs
pastProgressiveCount
1
totalVerbs
157
matches
0
"was leaving"
100.0000%
Purple prose (modifier overload)
Target: <4% adverbs, <2% -ly adverbs, no adj stacking
wordCount
955
adjectiveStacks
0
stackExamples
(empty)
adverbCount
37
adverbRatio
0.0387434554973822
lyAdverbCount
7
lyAdverbRatio
0.007329842931937173
100.0000%
Repeated phrase echo
Target: ≤20% sentences with echoes (window: 2)
totalSentences
109
echoCount
0
echoWords
(empty)
100.0000%
Sentence length variance
Target: CV ≥0.4 for sentence word counts
totalSentences
109
mean
10.07
std
6.34
cv
0.63
sampleLengths
0
11
1
23
2
20
3
8
4
5
5
2
6
4
7
17
8
13
9
5
10
4
11
10
12
17
13
24
14
10
15
8
16
22
17
14
18
2
19
3
20
10
21
2
22
10
23
4
24
7
25
5
26
32
27
1
28
13
29
17
30
8
31
5
32
10
33
12
34
11
35
29
36
4
37
17
38
10
39
20
40
11
41
8
42
7
43
9
44
4
45
4
46
15
47
3
48
12
49
6
82.8746%
Sentence opener variety
Target: ≥60% unique sentence openers
consecutiveRepeats
3
diversityRatio
0.5137614678899083
totalSentences
109
uniqueOpeners
56
75.0000%
Dialogue tag variety (said vs. fancy)
Target: ≤30% fancy dialogue tags
totalTags
5
fancyCount
2
fancyTags
0
"Patel said, circling to the other side (say, circle)"
1
"she said finally (say finally)"
90.6415%