NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-17-01KHMPADVYF381H8J68KXHXE5W
Run Details
Model:
DeepSeek-V2 Chat
Test:
Bad Writing Habits
Scenario:
Mystery: examining a crime scene
Duration:
40.8s
Cost:
$0.001840
Provider:
Novita
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 Tube station beneath Camden smelled of damp concrete and something older, something metallic that lingered at the back of Harlow Quinn's throat. Her boots
echoed
against cracked tiles as she stepped past the yellow police tape, the beam of her flashlight cutting through the gloom.
The air was thick with
the kind of quiet that only existed in places forgotten by time. "Detective." A uniformed officer nodded at her, his face pale under the flickering emergency lights. "Scene's just up ahead. DS
Evans
is already down there." Quinn adjusted the worn leather
watch
on her wrist—a nervous habit she'd picked up after Morris—and moved forward. The platform str
etched
before her, its edges swallowed by shadows. The remnants of the Veil Market were everywhere: overturned stalls, scattered trinkets, the faint scent of burnt sage. Someone had torn through this place like a storm.
Evans
stood near the center, his broad frame silhouetted against the glow of a single work lamp. He turned as she approached, his expression tight. "Took you long enough." "Traffic," Quinn lied. She'd been three blocks away when the call came in,
watch
ing a suspect's flat. The moment she'd heard the location, she'd known it would be trouble.
Evans
gestured to the body on the ground. "John Doe, mid-thirties. No wallet, no ID. Cause of death looks like blunt force trauma, but..." He trailed off, rubbing his jaw. Quinn crouched beside the corpse. The man's face was unmarked, but his shirt was soaked through with blood. She pulled on gloves and lifted the fabric
gently
. The wound beneath wasn't from any blunt object she'd ever seen—it was too clean, too precise, the edges blackened as if seared. "Knife?"
Evans
offered. Quinn shook her head. "No knife
does that." She
tilted her head, studying the pattern. The skin around the wound
was mottled
with faint, spiraling marks—sigils, maybe, though she didn't say that out loud. Not yet.
Evans
sighed. "Look, I know you've got your theories about this place, but let's not overcomplicate it. Gang hit, probably. The Veil Market's got plenty of enemies." Quinn didn't answer. Her gaze drifted to the ground near the body, where something glinted in the lamplight. A small brass compass, its casing weathered with verdigris. She picked it up
carefully
, turning it over in her palm. The face
was etched
with symbols she didn't recognize, the needle spinning
lazily
before settling—not north, but toward the far wall of the station.
Evans
frowned. "What's that?" "Evidence," she said, slipping it into an evidence bag before he could get a closer look. A noise
echoed
from the tunnels—a scrape, like metal on stone. Quinn's hand went to her sidearm.
Evans
stiffened beside her. "You hear that?" She nodded, already moving toward the sound, her flashlight cutting through the dark. The beam caught a figure darting behind a pillar. "Met Police!" Quinn called. "Show yourself!" Silence. Then, a whisper of movement. A young woman stepped into the light, her hands raised. Freckles dotted her nose, her curly red hair half-tucked behind one ear in a nervous gesture. Round glasses reflected the flashlight's glare. "I—I didn't mean to startle you," she stammered. Quinn kept her gun leveled. "Who are you?" "Eva Kowalski. I work at the British Museum. I was just—" Her gaze
flicked
to the body, then away. "Research."
Evans
snorted. "At a crime scene?" Eva swallowed. "I heard about the Market. I study occult history. I thought there might be... artifacts." Quinn's grip on the compass tightened. Too convenient. She studied Eva's face—the way her eyes kept darting to the compass, the way her fingers twitched at her sides. "You know what this is?" Eva hesitated. "It's a Veil Compass. It points to supernatural rifts."
Evans
rolled his eyes. "Christ. Not this again." Quinn ignored him. "And why would someone leave it here?" Eva's voice dropped. "Because they were looking for something. Or running from it." A chill prickled the back of Quinn's neck. She thought of Morris, of the case that had never made sense, of the way the evidence had twisted in on itself until she'd stopped trusting her own notes.
Evans
clapped his hands together. "Alright, enough ghost stories. Miss Kowalski, you're coming down to the station for questioning. And Quinn—" He shot her a look. "Don't go chasing shadows." Quinn
watch
ed as Eva
was led
away, the girl glancing back once, her green eyes wide with something like warning. The compass in Quinn's pocket felt heavier than it should have. She turned her flashlight toward the far wall, where the needle had pointed. The tiles there
were cracked
, the mortar between them blackened. And if she looked closely—if she
really
looked—she could almost see the faintest shimmer in the air, like heat off pavement in summer. The evidence didn't add up. But maybe, just maybe, it
was adding
up to something worse.
created_at
"2026-02-17T02:20:08.624212+00:00"
2/17/2026, 2:20:08 AM
model
"deepseek/deepseek-chat-v3"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1643
moderation_latency
(null)
generation_time
40393
tokens_prompt
923
tokens_completion
1300
native_tokens_prompt
856
native_tokens_completion
1152
native_tokens_completion_images
0
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
0.00184
router
(null)
provider_responses
0
id
"8f872501c48460d77bbf2033a1ef4829"
status
200
is_byok
false
latency
1643
endpoint_id
"cefb5a1e-9fea-4496-9a64-4e1b11a7cc8c"
provider_name
"Novita"
model_permaslug
"deepseek/deepseek-chat-v3"
api_type
"completions"
id
"gen-1771294767-R6lIDVsEOD6doLdHy7be"
upstream_id
"8f872501c48460d77bbf2033a1ef4829"
total_cost
0.00184
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Novita"
Evaluation details
Result
Evaluator
Details
Meta Data
100.0000%
Adverbs in dialogue tags
Target: ≤10% dialogue tags with adverbs
totalTags
4
adverbTagCount
0
adverbTags
(empty)
51.5152%
AI-ism adverb frequency
Target: <1% AI-ism adverbs (57 tracked)
wordCount
825
totalAiIsmAdverbs
4
found
0
adverb
"carefully"
count
1
1
adverb
"gently"
count
1
2
adverb
"lazily"
count
1
3
adverb
"really"
count
1
highlights
0
"carefully"
1
"gently"
2
"lazily"
3
"really"
80.0000%
AI-ism character names
Target: 0 AI-default names (17 tracked, −20% each)
codexExemptions
(empty)
found
0
"Evans"
100.0000%
AI-ism location names
Target: 0 AI-default location names (33 tracked, −20% each)
codexExemptions
(empty)
found
(empty)
51.5152%
AI-ism word frequency
Target: <1% AI-ism words (140 tracked)
wordCount
825
totalAiIsms
4
found
0
word
"flicked"
count
1
1
word
"echoed"
count
2
2
word
"etched"
count
1
highlights
0
"flicked"
1
"echoed"
2
"etched"
100.0000%
Cliche density
Target: ≤1 cliche(s) per 800-word window
totalCliches
1
maxInWindow
1
found
0
label
"air was thick with"
count
1
highlights
0
"The air was thick with"
100.0000%
Emotion telling (show vs. tell)
Target: ≤3% sentences with emotion telling
emotionTells
0
narrationSentences
66
matches
(empty)
100.0000%
Filter word density
Target: ≤12% sentences with filter/hedge words
filterCount
1
hedgeCount
0
narrationSentences
66
filterMatches
0
"watch"
hedgeMatches
(empty)
100.0000%
Overuse of "that" (subordinate clause padding)
Target: ≤10% sentences with "that" clauses
thatCount
1
totalSentences
82
matches
0
"does that.\" She"
100.0000%
Paragraph length variance
Target: CV ≥0.5 for paragraph word counts
totalParagraphs
32
mean
25.41
std
16.81
cv
0.662
sampleLengths
0
64
1
25
2
55
3
29
4
29
5
30
6
49
7
3
8
36
9
27
10
62
11
4
12
16
13
24
14
28
15
6
16
40
17
8
18
20
19
6
20
17
21
33
22
11
23
8
24
10
25
13
26
37
27
30
28
31
29
46
30
5
31
11
92.9293%
Passive voice overuse
Target: ≤5% passive sentences
passiveCount
4
totalSentences
66
matches
0
"was mottled"
1
"was etched"
2
"was led"
3
"were cracked"
100.0000%
Past progressive (was/were + -ing) overuse
Target: ≤10% past progressive verbs
pastProgressiveCount
1
totalVerbs
120
matches
0
"was adding"
100.0000%
Purple prose (modifier overload)
Target: <4% adverbs, <2% -ly adverbs, no adj stacking
wordCount
656
adjectiveStacks
0
stackExamples
(empty)
adverbCount
25
adverbRatio
0.038109756097560975
lyAdverbCount
7
lyAdverbRatio
0.010670731707317074
100.0000%
Repeated phrase echo
Target: ≤20% sentences with echoes (window: 2)
totalSentences
82
echoCount
0
echoWords
(empty)
100.0000%
Sentence length variance
Target: CV ≥0.4 for sentence word counts
totalSentences
82
mean
9.91
std
7.08
cv
0.714
sampleLengths
0
24
1
23
2
17
3
15
4
10
5
18
6
10
7
18
8
9
9
17
10
8
11
4
12
3
13
14
14
12
15
8
16
22
17
5
18
13
19
9
20
22
21
3
22
4
23
11
24
19
25
2
26
2
27
25
28
3
29
15
30
9
31
11
32
24
33
2
34
2
35
16
36
11
37
6
38
4
39
3
40
13
41
9
42
4
43
2
44
1
45
5
46
10
47
16
48
6
49
8
78.0488%
Sentence opener variety
Target: ≥60% unique sentence openers
consecutiveRepeats
1
diversityRatio
0.47560975609756095
totalSentences
82
uniqueOpeners
39
0.0000%
Dialogue tag variety (said vs. fancy)
Target: ≤30% fancy dialogue tags
totalTags
4
fancyCount
3
fancyTags
0
"Quinn lied (lie)"
1
"she said, slipping it into an evidence bag before he could get a closer look (say, slip get)"
2
"she stammered (stammer)"
85.5299%