NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-17-01KHMQXX8NSQ5GY40DPPSFA72V
Run Details
Model:
Claude Sonnet 4.5
Test:
Bad Writing Habits
Scenario:
Thriller: chase through city streets
Duration:
35.4s
Cost:
$0.025266
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="4417" name="Tomás Herrera" aliases="Tommy, Tomás, Herrera" pronouns="he/him" species="Human" storyRole="Support"> <eyeColor>warm brown</eyeColor> <hair>short curly dark brown</hair> <height>5'10"</height> <age>29</age> <skinTone>olive skin</skinTone> <scar>scar running along his left forearm from a knife attack</scar> <medallion>Saint Christopher medallion around his neck</medallion> <birthplace>Seville, Spain</birthplace> <description>A former paramedic who provides off-the-books medical care for the clique.</description> <backstory>Born in Seville, Spain. He moved to London to work for the NHS but lost his license after administering unauthorized treatments to supernatural patients.</backstory> </character> <location id="854" name="The Raven's Nest" aliases="the Nest, Raven, Silas' bar" realm="Earth"> <signColor>distinctive green neon sign above the entrance</signColor> <secretRoom>hidden back room accessible through a bookshelf, used for clandestine meetings</secretRoom> <description>A dimly lit bar in Soho, London. The walls are covered with old maps and black-and-white photographs.</description> </location> <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> </codex> Write a fully fleshed out scene of at least 1000 words in the thriller genre. A detective is pursuing a suspect through rain-soaked city streets at night. The chase takes them from the street level down into an underground market. The detective must decide whether to follow the suspect into unfamiliar and potentially dangerous territory. 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 rain hammered against the pavement like a thousand tiny fists, turning the Soho streets into mirrors of neon and shadow. Harlow Quinn's leather
watch
caught rainwater as she sprinted past The Raven's Nest, its distinctive green sign bleeding color into the downpour. Somewhere ahead in the maze of narrow lanes, Tomás Herrera's dark jacket flashed between pedestrians ducking under awnings. She'd been trailing him for three days,
watch
ing him slip in and out of that dingy bar, carrying medical bags at odd hours, meeting with known associates of the clique she'd been investigating for months. Tonight he'd spotted her. Rookie mistake—she'd gotten too close when he emerged from an alley behind a shuttered pharmacy, and their eyes had locked for one fatal second before he bolted. Her shoes pounded wet concrete, her breath coming sharp and measured. Military precision, even in pursuit. Her partner Morris had always teased her about that, the way she could maintain perfect form even when— She shoved the thought away. Morris
was gone
. Three years gone, and the case that took him still had holes in it she couldn't fill, questions that circled back on themselves like möbius strips. Herrera cut left down a service alley, his Saint Christopher medallion catching streetlight as he glanced back. The scar on his forearm was visible even at this distance when he raised his hand to vault a low railing. Former paramedic, her files said. Lost his NHS license under circumstances that remained frustratingly vague in the official records. But her unofficial sources—the ones who spoke in careful whispers at precinct bars—had painted a different picture. Unauthorized treatments. Patients who shouldn't have survived their injuries. Medical interventions that defied explanation. The same pattern that surrounded Morris's death. Quinn's jaw tightened as she followed him over the railing, landing hard on slick cobblestones. The alley opened onto a quieter street where Victorian buildings leaned together like old conspirators. Herrera was faster than she'd anticipated, his olive skin barely visible in the shadows between streetlamps. She pushed harder, her calves burning, rain streaming down her face and into her collar. He disappeared down a set of worn stone steps. Quinn slowed at the top, one hand on the rusted railing. The steps descended into darkness, spiraling down past the normal depth of a basement entrance. Water rushed down them in small cascades, and somewhere below came the
echo
of Herrera's
footsteps
—or something else entirely. She pulled out her mobile. No signal. Of course. The smart play was to call for backup. Station units at street level, wait for Herrera to emerge, bring him in for questioning at the precinct where she had procedures and protocols and witnesses. The safe play. The by-the-book play. But she'd been playing it safe for three years, and Morris's file remained stamped with that clinical word: unexplained. Quinn descended. The steps were older than they'd appeared from street level, the stone worn smooth by centuries of feet. She kept one hand on the damp wall, her other hand near her extendable baton. The sounds of the city above faded with each step—first the traffic, then the rain, until only her breathing and the drip of water accompanied her descent. The stairs ended in a narrow tunnel. Emergency lighting
flickered
overhead at irregular intervals, the kind installed in old Tube stations decades ago and never upgraded. But this wasn't on any Underground map she'd memorized. Camden was above her, she was certain of that much, but this tunnel ran perpendicular to the Northern Line. She moved forward, her brown eyes adjusting to the dimness. The tunnel's walls
were covered
in tiles that might have once been white but had aged to the color of old teeth. Water pooled in depressions on the floor, and she picked her way
carefully
around them, trying to keep her
footsteps
quiet. Voices drifted from ahead. Multiple voices, speaking in languages she didn't recognize. And something else—a smell that didn't belong in an abandoned Underground station. Incense, maybe, mixed with ozone and something organic and earthy that reminded her of the evidence locker where they kept confiscated drugs. The tunnel opened into a platform, and Quinn stopped dead. The space had been an Underground station once. She could see the bones of it—the curved ceiling, the tile work, the alcoves where advertisements would have hung. But it had
been transformed
into something else entirely. Market stalls crowded the platform, each one lit by lanterns that gave off light in colors that hurt to look at directly. People moved between them, and Quinn's mind stumbled over the word "people" because some of them weren't, couldn't be, shouldn't be— She forced her gaze back to ground level, to the concrete reality of stalls and merchandise and customers. Herrera stood thirty meters away, his back to her, speaking with a woman whose silver hair fell to her waist. Between them on a makeshift counter sat vials of liquid that
seemed to
contain small storms, glass bottles stopped with wax, bundles of dried herbs she didn't recognize. This was the Veil Market. She'd heard whispers about it from informants who'd later recanted their statements. An underground bazaar that moved locations, that sold things which couldn't be sold, that existed in the spaces between official reality. The same market Morris had mentioned the day before he died. Quinn's hand moved to her
watch
, a habit born of eighteen years on the force. The worn leather was familiar beneath her fingers, grounding. She had three choices. Go back up those stairs, file a report about what she'd seen, and
watch
it
get buried
like every other report that touched on these questions. Wait here and observe, gather intelligence, build a case that would inevitably fall apart when witnesses recanted and evidence disappeared. Or step onto that platform and demand answers from Herrera about what he was doing, who he was treating, and whether any of it connected to Morris. Thunder rumbled somewhere above, muffled by meters of London clay and Victorian engineering. The market sounds continued—bargaining, laughter, the clink of glass and metal. No one had noticed her yet. She was still an outsider, still had the advantage of anonymity. But Herrera was moving now, accepting a small package from the silver-haired woman, tucking it into his jacket. In seconds he'd disappear into the crowd or through one of the tunnel exits she could see at the far end of the platform. Quinn checked her
watch
one more time, felt the
weight
of her badge in her pocket, thought of Morris and those damned unexplained circumstances. Then she stepped onto the platform and into the market's strange light.
created_at
"2026-02-17T02:48:10.415519+00:00"
2/17/2026, 2:48:10 AM
model
"anthropic/claude-4.5-sonnet-20250929"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2695
moderation_latency
163
generation_time
35181
tokens_prompt
914
tokens_completion
1757
native_tokens_prompt
917
native_tokens_completion
1501
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
0.025266
router
(null)
provider_responses
0
id
"ec8fe255-3241-4340-baec-09fe78060ce3"
status
200
is_byok
false
latency
2694
endpoint_id
"d8bee848-f020-454a-b2be-4a7ce868f99f"
provider_name
"Amazon Bedrock"
model_permaslug
"anthropic/claude-4.5-sonnet-20250929"
api_type
"completions"
id
"gen-1771296453-c6qUROxiaFkwoonYgZ8h"
upstream_id
"ec8fe255-3241-4340-baec-09fe78060ce3"
total_cost
0.025266
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Amazon Bedrock"
Evaluation details
Result
Evaluator
Details
Meta Data
100.0000%
Adverbs in dialogue tags
Target: ≤10% dialogue tags with adverbs
totalTags
0
adverbTagCount
0
adverbTags
(empty)
91.0953%
AI-ism adverb frequency
Target: <1% AI-ism adverbs (57 tracked)
wordCount
1123
totalAiIsmAdverbs
1
found
0
adverb
"carefully"
count
1
highlights
0
"carefully"
100.0000%
AI-ism character names
Target: 0 AI-default names (17 tracked, −20% each)
codexExemptions
(empty)
found
(empty)
100.0000%
AI-ism location names
Target: 0 AI-default location names (33 tracked, −20% each)
codexExemptions
(empty)
found
(empty)
55.4764%
AI-ism word frequency
Target: <1% AI-ism words (140 tracked)
wordCount
1123
totalAiIsms
5
found
0
word
"flickered"
count
1
1
word
"weight"
count
1
2
word
"footsteps"
count
2
3
word
"echo"
count
1
highlights
0
"flickered"
1
"weight"
2
"footsteps"
3
"echo"
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
79
matches
(empty)
100.0000%
Filter word density
Target: ≤12% sentences with filter/hedge words
filterCount
3
hedgeCount
1
narrationSentences
79
filterMatches
0
"watch"
hedgeMatches
0
"seemed to"
100.0000%
Overuse of "that" (subordinate clause padding)
Target: ≤10% sentences with "that" clauses
thatCount
0
totalSentences
79
matches
(empty)
100.0000%
Paragraph length variance
Target: CV ≥0.5 for paragraph word counts
totalParagraphs
27
mean
41.19
std
26.05
cv
0.633
sampleLengths
0
61
1
66
2
34
3
34
4
88
5
7
6
61
7
9
8
45
9
9
10
40
11
19
12
2
13
60
14
54
15
53
16
46
17
10
18
79
19
66
20
38
21
11
22
101
23
41
24
42
25
24
26
12
99.5781%
Passive voice overuse
Target: ≤5% passive sentences
passiveCount
4
totalSentences
79
matches
0
"was gone"
1
"were covered"
2
"been transformed"
3
"get buried"
100.0000%
Past progressive (was/were + -ing) overuse
Target: ≤10% past progressive verbs
pastProgressiveCount
2
totalVerbs
184
matches
0
"was doing, was treating"
100.0000%
Purple prose (modifier overload)
Target: <4% adverbs, <2% -ly adverbs, no adj stacking
wordCount
1122
adjectiveStacks
0
stackExamples
(empty)
adverbCount
27
adverbRatio
0.02406417112299465
lyAdverbCount
8
lyAdverbRatio
0.0071301247771836
100.0000%
Repeated phrase echo
Target: ≤20% sentences with echoes (window: 2)
totalSentences
79
echoCount
0
echoWords
(empty)
100.0000%
Sentence length variance
Target: CV ≥0.4 for sentence word counts
totalSentences
79
mean
14.08
std
7.81
cv
0.555
sampleLengths
0
21
1
22
2
18
3
35
4
4
5
27
6
11
7
5
8
18
9
5
10
3
11
26
12
17
13
21
14
5
15
14
16
17
17
2
18
7
19
5
20
7
21
15
22
15
23
16
24
15
25
9
26
11
27
15
28
19
29
5
30
2
31
2
32
8
33
26
34
3
35
3
36
19
37
2
38
18
39
15
40
27
41
7
42
19
43
9
44
19
45
10
46
22
47
21
48
4
49
8
80.5907%
Sentence opener variety
Target: ≥60% unique sentence openers
consecutiveRepeats
3
diversityRatio
0.5063291139240507
totalSentences
79
uniqueOpeners
40
100.0000%
Dialogue tag variety (said vs. fancy)
Target: ≤30% fancy dialogue tags
totalTags
0
fancyCount
0
fancyTags
(empty)
95.6906%