NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYD4P93YA4KP3FA75J9PT6V2
Run Details
Model:
MoonshotAI: Kimi K3 (Reasoning, High)
Test:
Bad Writing Habits
Scenario:
Thriller: chase through city streets
Duration:
1m 17s
Cost:
$0.031440
Provider:
Moonshot AI
Evaluation summary
Score
Evaluator
Details Link
100.00%
Adverbs in dialogue tags
(details)
100.00%
AI-ism adverb frequency
(details)
100.00%
AI-ism character names
(details)
100.00%
AI-ism location names
(details)
85.40%
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)
100.00%
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)
100.00%
Passive voice overuse
(details)
95.29%
Past progressive (was/were + -ing) overuse
(details)
99.57%
Em-dash & semicolon overuse
(details)
100.00%
Purple prose (modifier overload)
(details)
100.00%
Repeated phrase echo
(details)
100.00%
Sentence length variance
(details)
56.23%
Sentence opener variety
(details)
100.00%
Adverb-first sentence starts
(details)
70.51%
Pronoun-first sentence starts
(details)
35.76%
Subject-first sentence starts
(details)
50.51%
Subordinate conjunction sentence starts
(details)
94.16%
Technical jargon density
(details)
100.00%
Useless dialogue additions
(details)
100.00%
Dialogue tag variety (said vs. fancy)
(details)
92.9138%
Overall
Model input
System
You are an expert fiction writer. Always keep the following rules in mind: - Write in past tense and use General English spelling, grammar, and colloquialisms/slang. - Write in active voice - Always follow the "show, don't tell" principle. - Avoid adverbs and cliches and overused/commonly used phrases. Aim for fresh and original descriptions. - Convey events and story through dialogue. - Mix short, punchy sentences with long, descriptive ones. Drop fill words to add variety. - Skip "he/she said said" dialogue tags and convey people's actions or face expressions through their speech - Avoid mushy dialog and descriptions, have dialogue always continue the action, never stall or add unnecessary fluff. Vary the descriptions to not repeat yourself. - Put dialogue on its own paragraph to separate scene and action. - Reduce indicators of uncertainty like "trying" or "maybe" When writing text: - NEVER conclude the scene on your own, follow the beat instructions very closely. - NEVER end with foreshadowing. - NEVER write further than what I prompt you with. - AVOID imagining possible endings, NEVER deviate from the instructions. - STOP EARLY if the continuation contains what was required in the instructions. You do not need to fill out the full amount of words possible.
User/Human
Take into account the following glossary of characters/locations/items/lore... when writing your response: <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> Continue the story using the following instructions: <instructions> 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. </instructions>
Model response
Messages
Meta
AI
The rain came down in
sheets that turned Soho's neon into smears of colour on the pavement.
Quinn's boots hammered the wet
stone, each stride closing the distance to the figure ahead of her. "Police! Stop!" The man didn't stop. They never did. He wore a grey hood, soaked through to a darker shade, and he moved like someone who knew these alleys better than the back of his own hands.
He cut left off Brewer
Street, down a passage barely wide enough for two people, and Quinn followed without breaking stride. Her shoulder clipped brick. She kept running. "You're making this worse!"
she shouted after him
. "We only want to talk!"
A lie, and they both
knew it.
She had a warrant in
her jacket pocket, laminated against the weather, with his face printed above a charge of conspiracy.
The same face she'd seen
twice before, leaving the Raven's Nest with people she'd been watching for six weeks.
The green neon of
Silas' bar still burned in her peripheral vision even here, three streets away, like an afterimage.
The man vaulted a wheelie
bin. She went through it instead, sending it clattering, rubbish fanning across the wet ground.
Her radio crackled at her
shoulder. "Quinn, what's your twenty? Over." She thumbed it without slowing. "On foot, pursuit, heading south through the alleys off Wardour. Male, grey hood, five-ten. I need units on the far side of—"
The man dropped out of
sight.
Quinn skidded to a halt
at the mouth of a stairwell, her boots slipping on the slick pavement.
A gap in the row
of shopfronts. A descending staircase she'd walked past a hundred times without noticing, squeezed between a shuttered kebab shop and a boarded-up pawnbroker's. No sign.
Just blackness going down, and
a worn handrail slick with rain.
She checked her watch out
of habit. 11:47. The leather strap hung loose where the third hole had worn through. "Control, suspect's gone underground. Some kind of stairwell between—"
She read the street numbers
. "Between 34 and 36 Beak Street. No visible signage." Static. Then: "Say again, Quinn? There's nothing between those numbers on our map."
She stared down into the
dark.
Somewhere below, footsteps echoed, quick
and unhurried both at once, the rhythm of a man who believed he'd already escaped. "He went down. I'm going after him." "Quinn, wait for backup. Quinn, do you copy?" She copied.
She also remembered Morris, three
years dead, and how they'd found him in a place that officially didn't exist, and how the report had said things she still couldn't square with the world.
She drew her torch and
her baton and went down the stairs. The rain's hiss faded behind her, replaced by a hum. Electrical, maybe. Or something else that lived in her teeth. The stairwell went deeper than any cellar had a right to go, curving, the walls shifting from brick to old yellow tile. Underground station tile. She
knew that tile
.
Her father had been a
Tube driver for thirty years. The stairs ended at a platform that shouldn't exist. Quinn stood at the threshold, torch raised, and forgot, for a moment, that she
was chasing
anyone. The platform stretched away in both directions, lit by lanterns strung on wire where the fluorescent tubes used to hang. Camden, she realised. This had to be one of the sealed branches beneath Camden, the ones closed before she was born. Except nobody had told the people here that it was closed. Stalls lined the walls, dozens of them, built from crates and salvaged doors and lengths of
velvet
. Candles burned in jars. The air smelled of damp stone and incense and something else, something sharp and herbal that made the back of her throat tighten. People filled the platform. Too many people for midnight, for a place that wasn't on any map. They browsed. They haggled. A woman with grey skin—grey, not pale, grey like wet cement—counted coins into a vendor's palm. A man in a three-piece suit examined a row of jars containing things that moved. Quinn lowered her torch. There, twenty metres down the platform, grey hood still up, moving through the crowd with the ease of a regular. Her suspect. He paused at a stall, said something to the vendor, kept walking. Her radio hissed. "Quinn, report. Quinn—" She killed the volume. Down here, the signal would be useless anyway. A man stepped into her path. Broad, watchful, with the stillness of hired security. He wore no uniform, but he didn't need one. "You lost, love?" "Metropolitan Police." She held up her warrant card. "I'm pursuing a suspect. Grey hood, came through not thirty seconds ahead of me." The man didn't look at the card. He looked at her face, then past her, up the stairs she'd descended,
as if measuring something
. "Entry's by token," he said. "You don't have a token." "I have a warrant." "That'll matter topside. Down here it matters less." He
wasn't threatening
her. That
unsettled
her more than a threat would have. "You've got two choices, Detective. Turn around and forget the stairs. Or walk in uninvited, in which case anything that happens is on you. House rules." "House rules." Her jaw tightened. "And whose house is this?" "It moves." He shrugged. "Full moon, it'll be somewhere else. You'd best decide quick." She looked past him. Her suspect had reached the far end of the platform, where the tunnel mouth gaped black. He glanced back.
Even at this distance, she
felt him
see
her.
Then he stepped into the
dark of the tunnel, and the crowd swallowed the space where he'd been. Eighteen years of service. Decorated. Commended twice. She had protocols for everything—armed suspects, hostage situations, bombs. Nothing in the manual covered a market in a ghost station where the doorman knew her rank before she said it and the patrons weren't entirely people. Morris had gone into a place like this alone. She'd always believed that. The case file called it a disused warehouse. The file
was written
by someone who hadn't seen the place, or someone who had and chose to lie. "Detective." The bouncer's voice had softened a fraction. "I've seen your sort come down before. The ones who turn back, we never
see
again. That's not a threat. That's just how it works." "And the ones who don't turn back?" He said nothing. Quinn thought about the warrant in her pocket, the hours of surveillance, the photographs pinned to the incident room board. The clique. Their meetings behind the bookshelf at the Raven's Nest. The paramedic who wasn't a paramedic anymore, patching wounds no hospital report ever recorded. Six weeks of work, and it all funnelled down to this: a threshold, a choice, a man in a grey hood vanishing into a tunnel. She clicked her torch back on. "I'm armed, and I'm going in after a suspect. If you want to obstruct a police investigation, you'll do it knowing exactly what that means." "Down here," the bouncer said, stepping aside, "it means nothing at all. That's the point." He studied her a moment longer, then reached into his coat and pressed something small and cold into her free hand. "You'll draw less attention holding this. Don't spend it." She looked down. A token, carved from bone, worn smooth at the edges. A face stared up from it. Not a face she recognised. Not a face she wanted to recognise. Quinn pocketed her baton, kept the torch low and angled at the ground, and walked into the market. The crowd parted around her without quite looking at her. She kept her eyes forward, her stride even, the way she'd learned to walk through prison corridors: like she belonged, like nothing here could touch her. Stalls passed on either side. Books bound in something that wasn't leather. Bottles that glowed faintly. A cage, small, ornate, with something inside that sang in a minor key. She did not stop. She did not stare. Ahead, the tunnel mouth waited, and somewhere inside it, her suspect
was still moving
, unhurried, certain she wouldn't follow. Quinn tightened her grip on the torch and walked into the dark.
created_at
"2026-07-25T17:20:43.657Z"
7/25/2026, 5:20:43 PM
model
"moonshotai/kimi-k3-20260715"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3987
moderation_latency
(null)
generation_time
76962
tokens_prompt
1104
tokens_completion
2121
native_tokens_prompt
1085
native_tokens_completion
1879
native_tokens_completion_images
(null)
native_tokens_reasoning
83
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
"stop"
service_tier
(null)
usage
0.03144
router
(null)
provider_responses
0
endpoint_id
"650c9b12-7b6d-444d-8d92-df8daef84d7a"
id
"chatcmpl-6a64f05da7f820f53971c40a"
is_byok
false
latency
3987
model_permaslug
"moonshotai/kimi-k3-20260715"
provider_name
"Moonshot AI"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.18.0; linux; x64))"
http_referer
(null)
request_id
"req-1785000043-J92HR50SJ9eShmwMOnUO"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1785000043-bGTrXlVkhvwnvPhiQ9OX"
upstream_id
"chatcmpl-6a64f05da7f820f53971c40a"
total_cost
0.03144
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Moonshot AI"
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
8
adverbTagCount
0
adverbTags
(empty)
dialogueSentences
29
tagDensity
0.276
leniency
0.552
rawRatio
0
effectiveRatio
0
100.00%
AI-ism adverb frequency
Target: <2% AI-ism adverbs (58 tracked)
wordCount
1370
totalAiIsmAdverbs
0
found
(empty)
highlights
(empty)
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)
85.40%
AI-ism word frequency
Target: <2% AI-ism words (290 tracked)
wordCount
1370
totalAiIsms
4
found
0
word
"footsteps"
count
1
1
word
"echoed"
count
1
2
word
"velvet"
count
1
3
word
"unsettled"
count
1
highlights
0
"footsteps"
1
"echoed"
2
"velvet"
3
"unsettled"
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
113
matches
(empty)
100.00%
Filter word density
Target: ≤3% sentences with filter/hedge words
filterCount
2
hedgeCount
0
narrationSentences
113
filterMatches
0
"watch"
1
"see"
hedgeMatches
(empty)
100.00%
Gibberish response detection
Target: ≤1% gibberish-like sentences (hard fail if a sentence exceeds 800 words)
analyzedSentences
132
gibberishSentences
0
adjustedGibberishSentences
0
longSentenceCount
0
runOnParagraphCount
0
giantParagraphCount
0
wordSaladCount
0
repetitionLoopCount
0
controlTokenCount
0
repeatedSegmentCount
0
maxSentenceWordsSeen
36
ratio
0
matches
(empty)
100.00%
Markdown formatting overuse
Target: ≤5% words in markdown formatting
markdownSpans
0
markdownWords
0
totalWords
1367
ratio
0
matches
(empty)
100.00%
Missing dialogue indicators (quotation marks)
Target: ≤10% speech attributions without quotation marks
totalAttributions
10
unquotedAttributions
0
matches
(empty)
100.00%
Name drop frequency
Target: ≤1.0 per-name mentions per 100 words
totalMentions
21
wordCount
1105
uniqueNames
10
maxNameDensity
0.72
worstName
"Quinn"
maxWindowNameDensity
1.5
worstWindowName
"Quinn"
discoveredNames
Soho
1
Brewer
1
Street
1
Quinn
8
Raven
2
Nest
2
Static
1
Morris
2
Tube
1
Camden
2
persons
0
"Quinn"
1
"Raven"
2
"Morris"
places
0
"Soho"
1
"Brewer"
2
"Street"
globalScore
1
windowScore
1
100.00%
Narrator intent-glossing
Target: ≤2% narration sentences with intent-glossing patterns
analyzedSentences
70
glossingSentenceCount
1
matches
0
"as if measuring something"
100.00%
"Not X but Y" pattern overuse
Target: ≤1 "not X but Y" per 1000 words
totalMatches
0
per1kWords
0
wordCount
1367
matches
(empty)
100.00%
Overuse of "that" (subordinate clause padding)
Target: ≤2% sentences with "that" clauses
thatCount
1
totalSentences
132
matches
0
"knew that tile"
100.00%
Paragraph length variance
Target: CV ≥0.5 for paragraph word counts
totalParagraphs
52
mean
26.29
std
21.58
cv
0.821
sampleLengths
0
34
1
2
2
7
3
56
4
13
5
67
6
26
7
5
8
27
9
6
10
59
11
20
12
23
13
13
14
26
15
7
16
8
17
47
18
59
19
9
20
17
21
96
22
52
23
4
24
34
25
6
26
12
27
23
28
3
29
22
30
24
31
10
32
4
33
47
34
10
35
14
36
50
37
43
38
40
39
33
40
7
41
3
42
70
43
6
44
25
45
45
46
31
47
18
48
65
49
8
100.00%
Passive voice overuse
Target: ≤2% passive sentences
passiveCount
1
totalSentences
113
matches
0
"was written"
95.29%
Past progressive (was/were + -ing) overuse
Target: ≤2% past progressive verbs
pastProgressiveCount
3
totalVerbs
191
matches
0
"was chasing"
1
"wasn't threatening"
2
"was still moving"
99.57%
Em-dash & semicolon overuse
Target: ≤2% sentences with em-dashes/semicolons
emDashCount
3
semicolonCount
0
flaggedSentences
2
totalSentences
132
ratio
0.015
matches
0
"A woman with grey skin—grey, not pale, grey like wet cement—counted coins into a vendor's palm."
1
"She had protocols for everything—armed suspects, hostage situations, bombs."
100.00%
Purple prose (modifier overload)
Target: <4% adverbs, <2% -ly adverbs, no adj stacking
wordCount
335
adjectiveStacks
0
stackExamples
(empty)
adverbCount
10
adverbRatio
0.029850746268656716
lyAdverbCount
2
lyAdverbRatio
0.005970149253731343
100.00%
Repeated phrase echo
Target: ≤20% sentences with echoes (window: 2)
totalSentences
132
echoCount
0
echoWords
(empty)
100.00%
Sentence length variance
Target: CV ≥0.4 for sentence word counts
totalSentences
132
mean
10.36
std
7.57
cv
0.731
sampleLengths
0
17
1
17
2
2
3
4
4
3
5
28
6
21
7
4
8
3
9
8
10
5
11
7
12
21
13
19
14
20
15
6
16
14
17
6
18
5
19
5
20
22
21
6
22
18
23
7
24
21
25
2
26
11
27
7
28
1
29
12
30
14
31
9
32
1
33
12
34
6
35
20
36
7
37
8
38
2
39
33
40
12
41
10
42
2
43
8
44
22
45
3
46
4
47
10
48
9
49
17
56.23%
Sentence opener variety
Target: ≥60% unique sentence openers
consecutiveRepeats
13
diversityRatio
0.3969465648854962
totalSentences
131
uniqueOpeners
52
100.00%
Adverb-first sentence starts
Target: ≥3% sentences starting with an adverb
adverbCount
4
totalSentences
99
matches
0
"Just blackness going down, and"
1
"Somewhere below, footsteps echoed, quick"
2
"Too many people for midnight,"
3
"Then he stepped into the"
ratio
0.04
70.51%
Pronoun-first sentence starts
Target: ≤30% sentences starting with a pronoun
pronounCount
37
totalSentences
99
matches
0
"They never did."
1
"He wore a grey hood,"
2
"He cut left off Brewer"
3
"Her shoulder clipped brick."
4
"She kept running."
5
"she shouted after him"
6
"She had a warrant in"
7
"She went through it instead,"
8
"Her radio crackled at her"
9
"She thumbed it without slowing."
10
"She checked her watch out"
11
"She read the street numbers"
12
"She stared down into the"
13
"She also remembered Morris, three"
14
"She drew her torch and"
15
"She knew that tile."
16
"Her father had been a"
17
"He paused at a stall,"
18
"Her radio hissed."
19
"She killed the volume."
ratio
0.374
35.76%
Subject-first sentence starts
Target: ≤72% sentences starting with a subject
subjectCount
84
totalSentences
99
matches
0
"The rain came down in"
1
"Quinn's boots hammered the wet"
2
"The man didn't stop."
3
"They never did."
4
"He wore a grey hood,"
5
"He cut left off Brewer"
6
"Her shoulder clipped brick."
7
"She kept running."
8
"she shouted after him"
9
"A lie, and they both"
10
"She had a warrant in"
11
"The same face she'd seen"
12
"The green neon of"
13
"The man vaulted a wheelie"
14
"She went through it instead,"
15
"Her radio crackled at her"
16
"She thumbed it without slowing."
17
"The man dropped out of"
18
"Quinn skidded to a halt"
19
"A gap in the row"
ratio
0.848
50.51%
Subordinate conjunction sentence starts
Target: ≥2% sentences starting with a subordinating conjunction
subConjCount
1
totalSentences
99
matches
0
"Even at this distance, she"
ratio
0.01
94.16%
Technical jargon density
Target: ≤6% sentences with technical-jargon patterns
analyzedSentences
44
technicalSentenceCount
3
matches
0
"Somewhere below, footsteps echoed, quick and unhurried both at once, the rhythm of a man who believed he'd already escaped."
1
"A man in a three-piece suit examined a row of jars containing things that moved."
2
"He looked at her face, then past her, up the stairs she'd descended, as if measuring something."
100.00%
Useless dialogue additions
Target: ≤5% dialogue tags with trailing filler fragments
totalTags
8
uselessAdditionCount
0
matches
(empty)
100.00%
Dialogue tag variety (said vs. fancy)
Target: ≤10% fancy dialogue tags
totalTags
3
fancyCount
1
fancyTags
0
"she shouted (shout)"
dialogueSentences
29
tagDensity
0.103
leniency
0.207
rawRatio
0.333
effectiveRatio
0.069
92.9138%