NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-17-01KHMMC98ESY6E4FC3JE3S39SG
Run Details
Model:
Claude 3.5 Sonnet
Test:
Bad Writing Habits
Scenario:
Mystery: examining a crime scene
Duration:
42.4s
Cost:
$0.046002
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
Detective Harlow Quinn crouched beside the body, ignoring the damp seeping through her trouser knees. The abandoned Tube platform hadn't seen regular commuters in decades, but tonight it hosted a grim tableau. The victim lay spread-eagled on the cracked tiles, limbs arranged with unnatural precision. No blood, no signs of struggle. Just a man in an expensive suit who looked like he'd decided to take a nap in a forgotten corner of London's underground. Except for the expression frozen on his face. Pure terror. Quinn checked her worn leather
watch
. 3:47 AM. The fluorescent work lights the forensics team had set up cast harsh shadows across the platform, turning the station's Victorian architectural details into leering faces. "Cause of death isn't immediately apparent," said Dr.
Chen
, the ME, snapping off her latex gloves. "No external trauma, no ligature marks. Tox screen pending, but-" "But you've never seen anything quite like it?" Quinn finished, her jaw tightening. The words were becoming too familiar lately. "Precisely."
Chen
gathered her equipment, clearly eager to get the body back to the morgue's better lighting and climate control. Quinn stood, her knees protesting, and surveyed the scene again. Something about this place needled at her consciousness. The platform showed signs of recent use - scuff marks on the tiles, modern graffiti layered over decades-old tags. But it wasn't the usual urban explorer territory. The access tunnel they'd used to reach this level had
been sealed
until someone reported strange noises two days ago. "Detective?" A familiar voice
echoed
across the platform. Eva Kowalski hurried toward her, red curls bouncing, her ever-present satchel slapping against her hip. The British Museum researcher had been consulting on several of Quinn's recent cases, though Quinn still wasn't entirely sure who had
authorized that arrangement
. "Ms. Kowalski. This is an active crime scene." Quinn's tone carried a warning, but Eva just pushed her round glasses up her nose and tucked a wayward strand of hair behind her left ear. "I know. That's why I'm here. You called about the symbols?" Quinn hadn't called anyone, but she'd learned to pick her battles. She gestured to the far wall where, half-hidden behind a modern advertisement poster, someone had carved an
intricate
pattern of interlocking circles and angular shapes. Eva's green
eyes widened
. She fumbled in her satchel, producing a small brass compass that looked antique but somehow wrong - too bright in places, too tarnished in others. Its face
was covered
in markings similar to the wall carvings. "Interesting," Eva murmured, holding the compass at arm's length. "This station... it's not on any official maps, is it?" "Technically, no." Quinn
watch
ed the younger woman
carefully
. "Care to share your thoughts?" Eva glanced at the body being loaded onto a gurney, then back at Quinn. "You've noticed there's no dust on him? In a sealed station?" Quinn had noticed. Just like she'd noticed how the temperature dropped ten degrees near the platform's north end, how their radio signals got staticky in specific spots, how the victim's expensive
watch
had stopped at exactly midnight despite the ME estimating time of death closer to 2 AM. "This isn't just any abandoned station," Eva continued, her freckled face intense. "This is one of the Market's old locations. See how the carvings form a perfect circle around the platform? Protection sigils. And these marks here?" She pointed to what Quinn had assumed was wear on the tile edge. "That's from bone tokens. Hundreds of them, over years." "The Market," Quinn repeated flatly. Another term she'd been hearing too often lately, always accompanied by impossible explanations for equally impossible crimes. Like what had
happened to
Morris three years ago... The compass in Eva's hand
suddenly
spun wildly, its needle whirling like a deranged weathervane. A cold draft swept the platform, sending paper evidence markers fluttering. "We should go," Eva said, shoving the compass back into her satchel. "The Market moves with the full moon, but it leaves...
echoes
. Especially when something's gone wrong." Quinn caught her arm. "Ms. Kowalski. I have a dead body and a crime scene that's making less sense by the minute. I need straight answers." "What you need," Eva said quietly, "is to start
accepting that some
crimes can't be solved with standard forensics and witness statements. Your victim didn't die from any cause Dr.
Chen
will find in her lab. He died from fear - pure, supernatural terror. And if we don't leave soon, we might find out exactly what scared him to death." The lights
flickered
. In the darkness between pulses, Quinn could have sworn she saw market stalls shimmer into existence along the platform, translucent figures browsing impossible wares. When the lights stabilized, there was only the empty station - but now she noticed the subtle patterns worn into the floor, exactly the width of vendor booths. "Fine," Quinn said, her military-precise posture tight with tension. "But tomorrow, you're going to explain everything you know about this Market. And Ms. Kowalski?" She fixed the researcher with a hard stare. "No more half-truths. I need to know what
really
happened to
my partner." Eva nodded, already hurrying toward the exit. "Tomorrow. I promise. But Detective?" She paused, looking back. "You might want to start carrying iron. Just in case." Quinn
watch
ed her go, then took one last look at the crime scene. In three years of investigating Morris's disappearance, she'd accumulated a file full of similar scenes - unexplainable deaths, impossible evidence, and always these strange symbols lurking at the edges. For the first time, she felt like she might be close to understanding what she was
really
investigating. She just wasn't sure she was ready for the answers.
created_at
"2026-02-17T01:46:13.313588+00:00"
2/17/2026, 1:46:13 AM
model
"anthropic/claude-3.5-sonnet"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1068
moderation_latency
135
generation_time
41835
tokens_prompt
923
tokens_completion
1512
native_tokens_prompt
937
native_tokens_completion
1346
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
0.046002
router
(null)
provider_responses
0
id
"af65151b-016d-43bd-98e2-6ae3d5c0e344"
status
200
is_byok
false
latency
1068
endpoint_id
"0e1957b3-205b-49d5-8413-13af3f68676d"
provider_name
"Amazon Bedrock"
model_permaslug
"anthropic/claude-3.5-sonnet"
api_type
"completions"
id
"gen-1771292730-TwwuKJHIbVyVmo7KypVz"
upstream_id
"af65151b-016d-43bd-98e2-6ae3d5c0e344"
total_cost
0.046002
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Amazon Bedrock"
Evaluation details
Result
Evaluator
Details
Meta Data
38.0952%
Adverbs in dialogue tags
Target: ≤10% dialogue tags with adverbs
totalTags
7
adverbTagCount
2
adverbTags
0
"Quinn repeated flatly [flatly]"
1
"Eva said quietly [quietly]"
47.2574%
AI-ism adverb frequency
Target: <1% AI-ism adverbs (57 tracked)
wordCount
948
totalAiIsmAdverbs
5
found
0
adverb
"carefully"
count
1
1
adverb
"precisely"
count
1
2
adverb
"suddenly"
count
1
3
adverb
"really"
count
2
highlights
0
"carefully"
1
"precisely"
2
"suddenly"
3
"really"
80.0000%
AI-ism character names
Target: 0 AI-default names (17 tracked, −20% each)
codexExemptions
(empty)
found
0
"Chen"
100.0000%
AI-ism location names
Target: 0 AI-default location names (33 tracked, −20% each)
codexExemptions
(empty)
found
(empty)
57.8059%
AI-ism word frequency
Target: <1% AI-ism words (140 tracked)
wordCount
948
totalAiIsms
4
found
0
word
"flickered"
count
1
1
word
"intricate"
count
1
2
word
"echoed"
count
1
3
word
"echoes"
count
1
highlights
0
"flickered"
1
"intricate"
2
"echoed"
3
"echoes"
100.0000%
Cliche density
Target: ≤1 cliche(s) per 800-word window
totalCliches
1
maxInWindow
1
found
0
label
"eyes widened/narrowed"
count
1
highlights
0
"eyes widened"
100.0000%
Emotion telling (show vs. tell)
Target: ≤3% sentences with emotion telling
emotionTells
0
narrationSentences
55
matches
(empty)
100.0000%
Filter word density
Target: ≤12% sentences with filter/hedge words
filterCount
1
hedgeCount
1
narrationSentences
55
filterMatches
0
"watch"
hedgeMatches
0
"happened to"
100.0000%
Overuse of "that" (subordinate clause padding)
Target: ≤10% sentences with "that" clauses
thatCount
2
totalSentences
65
matches
0
"authorized that arrangement"
1
"accepting that some"
100.0000%
Paragraph length variance
Target: CV ≥0.5 for paragraph word counts
totalParagraphs
27
mean
35.07
std
17.84
cv
0.509
sampleLengths
0
74
1
10
2
33
3
26
4
20
5
20
6
65
7
47
8
34
9
11
10
36
11
40
12
19
13
13
14
25
15
48
16
59
17
31
18
26
19
28
20
26
21
60
22
55
23
45
24
26
25
60
26
10
100.0000%
Passive voice overuse
Target: ≤5% passive sentences
passiveCount
2
totalSentences
55
matches
0
"been sealed"
1
"was covered"
100.0000%
Past progressive (was/were + -ing) overuse
Target: ≤10% past progressive verbs
pastProgressiveCount
0
totalVerbs
125
matches
(empty)
85.6425%
Purple prose (modifier overload)
Target: <4% adverbs, <2% -ly adverbs, no adj stacking
wordCount
690
adjectiveStacks
1
stackExamples
0
"over decades-old tags."
adverbCount
34
adverbRatio
0.04927536231884058
lyAdverbCount
14
lyAdverbRatio
0.020289855072463767
100.0000%
Repeated phrase echo
Target: ≤20% sentences with echoes (window: 2)
totalSentences
65
echoCount
0
echoWords
(empty)
100.0000%
Sentence length variance
Target: CV ≥0.4 for sentence word counts
totalSentences
65
mean
14.57
std
10.39
cv
0.713
sampleLengths
0
15
1
17
2
13
3
6
4
23
5
8
6
2
7
6
8
2
9
25
10
8
11
8
12
10
13
13
14
7
15
20
16
10
17
8
18
19
19
8
20
20
21
8
22
15
23
24
24
34
25
11
26
11
27
25
28
4
29
25
30
11
31
9
32
10
33
8
34
5
35
14
36
11
37
3
38
45
39
12
40
38
41
9
42
5
43
17
44
9
45
15
46
11
47
12
48
16
49
4
84.1026%
Sentence opener variety
Target: ≥60% unique sentence openers
consecutiveRepeats
2
diversityRatio
0.5230769230769231
totalSentences
65
uniqueOpeners
34
0.0000%
Dialogue tag variety (said vs. fancy)
Target: ≤30% fancy dialogue tags
totalTags
7
fancyCount
5
fancyTags
0
"Eva murmured, holding the compass at arm's length (murmur, hold)"
1
"Eva continued, her freckled face intense (continue)"
2
"Quinn repeated flatly (repeat flatly)"
3
"Eva said, shoving the compass back into her satchel (say, shove back)"
4
"Eva said quietly (say quietly)"
81.9355%