NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-20-01KHY95BSPXEK6NBBMBDPNDZBR
Run Details
Model:
Gemma 3 4B
Test:
Bad Writing Habits
Scenario:
Literary fiction: old friends reunite
Duration:
22.2s
Cost:
$0.000174
Provider:
DeepInfra
Evaluation summary
Score
Evaluator
Details Link
27.59%
Adverbs in dialogue tags
(details)
58.61%
AI-ism adverb frequency
(details)
100.00%
AI-ism character names
(details)
100.00%
AI-ism location names
(details)
17.22%
AI-ism word frequency
(details)
66.67%
Cliché density
(details)
100.00%
Emotion telling (show vs. tell)
(details)
90.59%
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)
0.00%
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)
100.00%
Past progressive (was/were + -ing) overuse
(details)
0.00%
Em-dash & semicolon overuse
(details)
96.08%
Purple prose (modifier overload)
(details)
100.00%
Repeated phrase echo
(details)
100.00%
Sentence length variance
(details)
56.31%
Sentence opener variety
(details)
100.00%
Adverb-first sentence starts
(details)
0.00%
Pronoun-first sentence starts
(details)
6.05%
Subject-first sentence starts
(details)
0.00%
Subordinate conjunction sentence starts
(details)
100.00%
Technical jargon density
(details)
0.00%
Useless dialogue additions
(details)
70.00%
Dialogue tag variety (said vs. fancy)
(details)
69.6372%
Overall
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="4412" name="Aurora Carter" aliases="Rory, Laila, Aurora, Carter, Malphora" pronouns="she/her" species="Human" storyRole="Protagonist"> <eyeColor>bright blue</eyeColor> <hair>straight shoulder-length black</hair> <height>5'6"</height> <age>25</age> <scar>small crescent-shaped scar on her left wrist from a childhood accident</scar> <description>Aurora, referred to by her friends as Rory, is cool-headed, intelligent, and known for quick out-of-the-box thinking.</description> <backstory>Born in Cardiff to an Irish barrister father (Brendan Carter) and Welsh teacher mother (Jennifer Carter née Ellis), Rory reluctantly pursued Pre-Law at Cardiff University. She fled to London at the behest of her childhood friend Eva after leaving an abusive ex named Evan. By day, Rory works part-time as a delivery person for Yu-Fei Cheung's Golden Empress restaurant. She lives in a flat above Silas' bar.</backstory> </character> <character id="4413" name="Silas Blackwood" aliases="Silas, Si, the Spymaster" pronouns="he/him" species="Human" storyRole="Mentor"> <eyeColor>hazel</eyeColor> <hair>grey-streaked auburn</hair> <height>6'1"</height> <age>58</age> <beard>neatly trimmed, matching his grey-streaked auburn hair</beard> <limp>slight limp in his left leg from an old knee injury</limp> <signetRing>silver signet ring, always worn on his right hand</signetRing> <description>Silas is a retired intelligence operative turned bar owner. He carries himself with quiet authority.</description> <backstory>Former MI6 field agent who retired after a botched operation in Prague left him with a permanent knee injury. He opened "The Raven's Nest" bar in Soho as a front for his network of contacts.</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> </codex> Write a fully fleshed out scene of at least 1000 words in the literary fiction genre. Two old friends who haven't seen each other in years run into each other unexpectedly at a bar. One of them has changed significantly. Through their conversation and the tension between who they were and who they've become, reveal the weight of time and unspoken regret. 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 hadn’t let up.
It hammered against the windows
of The Raven’s Nest, a relentless, grey
drumming that seemed to mirror the anxiety knotting in Aurora’s stomach
. She’d ordered a scotch, neat, and
was nursing
it
slowly
, watching the flickering gas lamps cast elongated shadows across the crowded bar.
Silas’s place was always a
little unsettling, a
repository
of hushed conversations and unsettlingly knowing glances. It smelled of aged wood, spilled beer, and something vaguely metallic – the ghost of a thousand secrets. Then she saw him.
He was standing near the
bar, nursing a whiskey, his back to her. Tall, still – a solid six-one, just like she remembered – but… different.
The easy, careless confidence of
Evan had
been replaced
with a guarded stillness.
He wore a charcoal grey
suit, impeccably tailored, and a silver signet ring flashed briefly as he shifted his
weight
.
It was the ring she
remembered from countless photographs, the one her father had insisted on photographing for a family album. It was Liam.
Time seemed to compress, the
sounds of the bar fading into a dull hum. Aurora set down her glass, the ice clinking
softly
against the
crystal
.
She hadn't spoken to him
in nearly seven years, not since she’d fled London with Eva, abandoning a life she hadn't
truly
wanted and, she now suspected, a friendship he hadn’t understood.
He turned then, his hazel
eyes
scanning
the room, lingering on her for a moment before registering recognition. A
flicker
of something – surprise? Regret? – crossed his face before he straightened, offering a curt nod. “Aurora,”
he said, the voice deeper
than she recalled, roughened by years of something she couldn’t
quite guess
. “Liam,”
she replied, her own voice
sounding flat, unfamiliar even to her. “It’s… been a while.” “Quite,”
he said, pulling up a
stool beside her.
The movement caused a slight
hitch in his left leg, a small, almost imperceptible limp that confirmed the old injury Silas had alluded to. “You look… well.” “Trying to,”
she said, taking a sip
of her scotch.
It burned a welcome path
down her throat. "You too. The suit's a bit… formal, isn't it?"
He chuckled, a dry, brittle
sound. “Business calls. I’ve moved on from casual Fridays.”
He paused, studying her
. “You haven’t changed much. Still got that bright blue stare.” It was a backhanded compliment, she realized, and she didn’t bother to hide her own
flicker
of defensiveness. “And you’re still devastatingly handsome, even with the grey creeping in.” The
silence
stretched between them, thick with
unspoken
history. The rain continued its insistent argument with the windows. Finally, Liam broke it. “So,”
he said, gesturing vaguely at
the room. “What do you do these days?” “Delivery,”
she answered, the words feeling
inadequate, a pale
echo
of her life. “Yu-Fei Cheung’s Golden Empress. It’s honest work. Keeps me busy.” He raised an
eyebrow
. “Delivery? Really?” “Someone’s got to get the dumplings to the people,”
she said, a touch of irony in her voice
. “It pays the rent.” “Evan always said you were… resourceful,” he commented, the words laced with a subtle disapproval that stung more than she’d expected. The name
hung in the air
, a loaded trigger. She’d
deliberately
avoided thinking about him, about the wreckage he’d left behind. “Evan is… a distant memory,”
she said, her voice clipped
. “Some memories don’t fade, Aurora,” he said quietly. “They just… get buried. Like things in a dustbin.” He had a way of looking at her, a penetrating
gaze that seemed to strip away layers of pretense
. She looked away, focusing on the swirling amber liquid in her glass. “I left,” she said, the words a hesitant plea for justification. “You were… difficult.” “Difficult is a polite way of putting it,” he acknowledged. “You walked away without a word. Without a goodbye.” “I had to,”
she said, the irritation rising
. “It wasn’t safe.” “Safe isn’t always what’s important,”
he said, his voice low
. “Sometimes, the bravest thing you can do is face the storm.” “I was seventeen,”
she countered, the frustration bubbling over
. “I didn’t know what I was doing.” “No one does at seventeen,” he agreed, his eyes holding a trace of something she couldn't
quite name – sadness, perhaps, or regret
. “I should have been the one to tell you. I should have… done something.” “You didn’t,”
she said, the words coming out sharper than she intended
. He didn’t argue. “I was… compromised,” he offered, a vague explanation that didn’t
quite satisfy
. “Let’s just say I had my own problems.” He took a long swallow of his whiskey. The neon green sign of The Raven’s Nest
pulsed
overhead, casting an eerie glow on his face. “I’ve been… busy,” he said. "Working. Some things don't get better with time, Aurora. They just… harden." She looked at him,
really
looked at him, and saw the lines
etched
around his eyes, the weariness in his posture. She saw the ghost of the boy she’d known, the one who’d shared secrets and dreams in a thousand smoky rooms. But beneath it all, there was something else – a
profound
sense of
disillusionment, a quiet acceptance of a life lived in the shadows. "What kind of work?"
she asked, the question deliberately neutral
. He hesitated, then said, “Let’s just say I deal in information. Discreetly. The kind people prefer to keep hidden. I retired from MI6 a few years back. The Prague operation... it left its mark." The words
hung in the air
, confirming rumors she’d long dismissed as the fanciful exaggerations of late-night bar conversations. Silas had mentioned a botched operation, a
shattered
knee – but she’d never connected it to Liam. “Silas knows about that,”
she said, a sudden chill running through her
. He nodded, a grim acknowledgement. “He’s a good man, Silas. A
very
good man. He offered me a place here, at The Raven’s Nest. A quiet corner to observe, to listen.” He gestured around the bar, at the worn faces and hushed tones. “A place to forget.” “You haven’t forgotten,” she said, stating the obvious. “Some things… you can’t quite erase,” he admitted. “Like the feeling of rain on your face after a long
silence
.” He paused, then added, almost as an afterthought, “I often wondered if you were happy, Aurora. If you’d found a life that suited you.” She considered her answer
carefully
. The life she had was… a refuge. A simple,
solitary
existence that shielded her from the wreckage of the past. It wasn’t glamorous, it wasn’t exciting, but it was safe. “I’m content,” she said finally, the word a small, fragile offering. Liam studied her face, his hazel eyes searching for any hint of deception. He didn’t press her.
Instead, he lifted his glass
in a silent toast. “To contentment,”
he said, his voice barely audible above the drumming rain
. “And to the ghosts we leave behind.” She raised her own glass in return, the ice clinking
softly
against the
crystal
. The rain kept falling, washing the grime from the streets of London, and blurring the edges of their shared history. She knew, with a certainty that settled deep in her bones, that this was the last time they would see each other. And perhaps, she realized, it was for the best.
created_at
"2026-02-20T19:41:53.854Z"
2/20/2026, 7:41:53 PM
model
"google/gemma-3-4b-it"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
262
moderation_latency
(null)
generation_time
22040
tokens_prompt
892
tokens_completion
1985
native_tokens_prompt
850
native_tokens_completion
1754
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
0.00017432
router
(null)
provider_responses
0
endpoint_id
"d3de3bd4-81bc-48fb-924f-2a87b2a36e75"
id
"chatcmpl-RceZeRDvamnIvx09dCKFpiLl"
is_byok
false
latency
125
model_permaslug
"google/gemma-3-4b-it"
provider_name
"DeepInfra"
status
200
api_type
"completions"
id
"gen-1771616513-seCfo1R51w7BY9anGIWG"
upstream_id
"chatcmpl-RceZeRDvamnIvx09dCKFpiLl"
total_cost
0.00017432
cache_discount
(null)
upstream_inference_cost
0
provider_name
"DeepInfra"
Evaluation details
Result
Evaluator
Details
Meta Data
27.59%
Adverbs in dialogue tags
Target: ≤10% dialogue tags with adverbs
totalTags
29
adverbTagCount
5
adverbTags
0
"he said quietly [quietly]"
1
"He hesitated then [then]"
2
"He gestured around [around]"
3
"He paused then [then]"
4
"she said finally [finally]"
dialogueSentences
50
tagDensity
0.58
leniency
1
rawRatio
0.172
effectiveRatio
0.172
58.61%
AI-ism adverb frequency
Target: <2% AI-ism adverbs (58 tracked)
wordCount
1208
totalAiIsmAdverbs
10
found
0
adverb
"slowly"
count
1
1
adverb
"softly"
count
2
2
adverb
"truly"
count
1
3
adverb
"really"
count
2
4
adverb
"deliberately"
count
2
5
adverb
"very"
count
1
6
adverb
"carefully"
count
1
highlights
0
"slowly"
1
"softly"
2
"truly"
3
"really"
4
"deliberately"
5
"very"
6
"carefully"
100.00%
AI-ism character names
Target: 0 AI-default names (16 tracked, −20% each)
codexExemptions
0
"Blackwood"
found
(empty)
100.00%
AI-ism location names
Target: 0 AI-default location names (33 tracked, −20% each)
codexExemptions
(empty)
found
(empty)
17.22%
AI-ism word frequency
Target: <2% AI-ism words (290 tracked)
wordCount
1208
totalAiIsms
20
found
0
word
"stomach"
count
1
1
word
"repository"
count
1
2
word
"weight"
count
1
3
word
"crystal"
count
2
4
word
"scanning"
count
1
5
word
"flicker"
count
2
6
word
"silence"
count
2
7
word
"unspoken"
count
1
8
word
"echo"
count
1
9
word
"eyebrow"
count
1
10
word
"pulsed"
count
1
11
word
"etched"
count
1
12
word
"profound"
count
1
13
word
"sense of"
count
1
14
word
"shattered"
count
1
15
word
"chill"
count
1
16
word
"solitary"
count
1
highlights
0
"stomach"
1
"repository"
2
"weight"
3
"crystal"
4
"scanning"
5
"flicker"
6
"silence"
7
"unspoken"
8
"echo"
9
"eyebrow"
10
"pulsed"
11
"etched"
12
"profound"
13
"sense of"
14
"shattered"
15
"chill"
16
"solitary"
66.67%
Cliché density
Target: ≤1 cliche(s) per 800-word window
totalCliches
2
maxInWindow
2
found
0
label
"hung in the air"
count
2
highlights
0
"hung in the air"
100.00%
Emotion telling (show vs. tell)
Target: ≤3% sentences with emotion telling
emotionTells
0
narrationSentences
82
matches
(empty)
90.59%
Filter word density
Target: ≤3% sentences with filter/hedge words
filterCount
0
hedgeCount
3
narrationSentences
82
filterMatches
(empty)
hedgeMatches
0
"seemed to"
100.00%
Gibberish response detection
Target: ≤1% gibberish-like sentences (hard fail if a sentence exceeds 800 words)
analyzedSentences
103
gibberishSentences
0
adjustedGibberishSentences
0
longSentenceCount
0
runOnParagraphCount
0
giantParagraphCount
0
wordSaladCount
0
repetitionLoopCount
0
controlTokenCount
0
maxSentenceWordsSeen
38
ratio
0
matches
(empty)
100.00%
Markdown formatting overuse
Target: ≤5% words in markdown formatting
markdownSpans
0
markdownWords
0
totalWords
1214
ratio
0
matches
(empty)
100.00%
Missing dialogue indicators (quotation marks)
Target: ≤10% speech attributions without quotation marks
totalAttributions
34
unquotedAttributions
0
matches
(empty)
100.00%
Name drop frequency
Target: ≤1.0 per-name mentions per 100 words
totalMentions
17
wordCount
889
uniqueNames
8
maxNameDensity
0.45
worstName
"Liam"
maxWindowNameDensity
1
worstWindowName
"Aurora"
discoveredNames
Raven
2
Nest
2
Aurora
2
Evan
1
Liam
4
London
2
Eva
1
Silas
3
persons
0
"Raven"
1
"Aurora"
2
"Evan"
3
"Liam"
4
"Eva"
5
"Silas"
places
0
"London"
globalScore
1
windowScore
1
0.00%
Narrator intent-glossing
Target: ≤2% narration sentences with intent-glossing patterns
analyzedSentences
59
glossingSentenceCount
5
matches
0
"drumming that seemed to mirror the anxiety knotting in Aurora’s stomach"
1
"quite guess"
2
"gaze that seemed to strip away layers of pretense"
3
"quite name – sadness, perhaps, or regret"
4
"quite satisfy"
100.00%
"Not X but Y" pattern overuse
Target: ≤1 "not X but Y" per 1000 words
totalMatches
0
per1kWords
0
wordCount
1214
matches
(empty)
100.00%
Overuse of "that" (subordinate clause padding)
Target: ≤2% sentences with "that" clauses
thatCount
0
totalSentences
103
matches
(empty)
100.00%
Paragraph length variance
Target: CV ≥0.5 for paragraph word counts
totalParagraphs
43
mean
28.23
std
18.95
cv
0.671
sampleLengths
0
84
1
4
2
80
3
3
4
58
5
36
6
18
7
16
8
36
9
27
10
28
11
29
12
22
13
14
14
23
15
6
16
22
17
21
18
31
19
17
20
45
21
19
22
11
23
21
24
16
25
37
26
12
27
23
28
42
29
66
30
10
31
34
32
36
33
12
34
47
35
8
36
20
37
24
38
35
39
11
40
26
41
19
42
65
100.00%
Passive voice overuse
Target: ≤2% passive sentences
passiveCount
1
totalSentences
82
matches
0
"been replaced"
100.00%
Past progressive (was/were + -ing) overuse
Target: ≤2% past progressive verbs
pastProgressiveCount
2
totalVerbs
162
matches
0
"was nursing"
1
"was standing"
0.00%
Em-dash & semicolon overuse
Target: ≤2% sentences with em-dashes/semicolons
emDashCount
8
semicolonCount
0
flaggedSentences
7
totalSentences
103
ratio
0.068
matches
0
"It smelled of aged wood, spilled beer, and something vaguely metallic – the ghost of a thousand secrets."
1
"Tall, still – a solid six-one, just like she remembered – but…"
2
"A flicker of something – surprise?"
3
"– crossed his face before he straightened, offering a curt nod."
4
"“No one does at seventeen,” he agreed, his eyes holding a trace of something she couldn't quite name – sadness, perhaps, or regret."
5
"But beneath it all, there was something else – a profound sense of disillusionment, a quiet acceptance of a life lived in the shadows."
6
"Silas had mentioned a botched operation, a shattered knee – but she’d never connected it to Liam."
96.08%
Purple prose (modifier overload)
Target: <4% adverbs, <2% -ly adverbs, no adj stacking
wordCount
883
adjectiveStacks
0
stackExamples
(empty)
adverbCount
38
adverbRatio
0.04303510758776897
lyAdverbCount
19
lyAdverbRatio
0.021517553793884484
100.00%
Repeated phrase echo
Target: ≤20% sentences with echoes (window: 2)
totalSentences
103
echoCount
0
echoWords
(empty)
100.00%
Sentence length variance
Target: CV ≥0.4 for sentence word counts
totalSentences
103
mean
11.79
std
7.1
cv
0.603
sampleLengths
0
5
1
23
2
22
3
16
4
18
5
4
6
13
7
12
8
1
9
13
10
20
11
21
12
3
13
14
14
12
15
32
16
18
17
6
18
1
19
11
20
18
21
12
22
4
23
9
24
24
25
3
26
10
27
8
28
9
29
6
30
12
31
10
32
18
33
11
34
9
35
9
36
4
37
8
38
6
39
13
40
10
41
4
42
2
43
18
44
4
45
21
46
9
47
12
48
10
49
8
56.31%
Sentence opener variety
Target: ≥60% unique sentence openers
consecutiveRepeats
7
diversityRatio
0.3786407766990291
totalSentences
103
uniqueOpeners
39
100.00%
Adverb-first sentence starts
Target: ≥3% sentences starting with an adverb
adverbCount
3
totalSentences
76
matches
0
"Then she saw him."
1
"Finally, Liam broke it."
2
"Instead, he lifted his glass"
ratio
0.039
0.00%
Pronoun-first sentence starts
Target: ≤30% sentences starting with a pronoun
pronounCount
52
totalSentences
76
matches
0
"It hammered against the windows"
1
"She’d ordered a scotch, neat,"
2
"It smelled of aged wood,"
3
"He was standing near the"
4
"He wore a charcoal grey"
5
"It was the ring she"
6
"It was Liam."
7
"She hadn't spoken to him"
8
"He turned then, his hazel"
9
"he said, the voice deeper"
10
"she replied, her own voice"
11
"he said, pulling up a"
12
"she said, taking a sip"
13
"It burned a welcome path"
14
"He chuckled, a dry, brittle"
15
"He paused, studying her"
16
"It was a backhanded compliment,"
17
"he said, gesturing vaguely at"
18
"she answered, the words feeling"
19
"He raised an eyebrow."
ratio
0.684
6.05%
Subject-first sentence starts
Target: ≤72% sentences starting with a subject
subjectCount
69
totalSentences
76
matches
0
"The rain hadn’t let up."
1
"It hammered against the windows"
2
"She’d ordered a scotch, neat,"
3
"Silas’s place was always a"
4
"It smelled of aged wood,"
5
"He was standing near the"
6
"The easy, careless confidence of"
7
"He wore a charcoal grey"
8
"It was the ring she"
9
"It was Liam."
10
"Time seemed to compress, the"
11
"Aurora set down her glass,"
12
"She hadn't spoken to him"
13
"He turned then, his hazel"
14
"A flicker of something –"
15
"he said, the voice deeper"
16
"she replied, her own voice"
17
"he said, pulling up a"
18
"The movement caused a slight"
19
"she said, taking a sip"
ratio
0.908
0.00%
Subordinate conjunction sentence starts
Target: ≥2% sentences starting with a subordinating conjunction
subConjCount
0
totalSentences
76
matches
(empty)
ratio
0
100.00%
Technical jargon density
Target: ≤6% sentences with technical-jargon patterns
analyzedSentences
31
technicalSentenceCount
1
matches
0
"It hammered against the windows of The Raven’s Nest, a relentless, grey drumming that seemed to mirror the anxiety knotting in Aurora’s stomach."
0.00%
Useless dialogue additions
Target: ≤5% dialogue tags with trailing filler fragments
totalTags
29
uselessAdditionCount
11
matches
0
"she replied, her own voice sounding flat, unfamiliar even to her"
1
"she answered, the words feeling inadequate, a pale echo of her life"
2
"she said, a touch of irony in her voice"
3
"she said, her voice clipped"
4
"she said, the irritation rising"
5
"he said, his voice low"
6
"she countered, the frustration bubbling over"
7
"she said, the words coming out sharper than she intended"
8
"she asked, the question deliberately neutral"
9
"she said, a sudden chill running through her"
10
"he said, his voice barely audible above the drumming rain"
70.00%
Dialogue tag variety (said vs. fancy)
Target: ≤10% fancy dialogue tags
totalTags
23
fancyCount
4
fancyTags
0
"he commented (comment)"
1
"he acknowledged (acknowledge)"
2
"he agreed (agree)"
3
"he admitted (admit)"
dialogueSentences
50
tagDensity
0.46
leniency
0.92
rawRatio
0.174
effectiveRatio
0.16
69.6372%