NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1KATB9CTTKZ5TVWYFRM4516
Run Details
Model:
DeepSeek V4 Flash 0731 (Reasoning, High)
Test:
Language Writing
Scenario:
Character dialogue (Spanish) in a story
Duration:
1m 8s
Cost:
$0.000341
Provider:
OpenInference
Evaluation summary
Score
Evaluator
Details Link
37.50%
Parse dialogue
(details)
Model input
System
You are an expert novel author in the cozy mystery genre. Take the following details about characters/locations/... into account: <character name="Maria"> Maria is a 30-year old college student on a study abroad program. Born in Spain, she is fluent in Spanish and does not yet know any English. She enjoys landscape photography and is always seen with her camera. </character> <character name="Hunter"> Hunter is a 35-year old detective who is investigating a series of mysterious disappearances in the small town of Alborn, MN. He is a former marine and has a no-nonsense attitude. Hunter has no knowledge of Spanish, not even basic phrases. </character> <location name="Alborn, MN"> Alborne is a small town in the Midwest with a population of 3,000. It is known for its annual apple festival. </location>
User/Human
Write the following scene: Maria is sitting at a cafe, looking at her camera. Hunter approaches her and asks her about a recent event in the town. She responds troubled and shaken in Spanish, and Hunter looks confused. They talk for a while, until Hunter gets a call and has to leave. Return the dialogue between Maria and Hunter as a voice sheet: Maria: "..." Hunter: "..."
Model response
Messages
Meta
AI
The late afternoon sun cast long shadows across the cracked pavement of Main Street, filtering through the dusty window of "The Rusty Mug." Inside, the air was thick with the scent of old coffee and simmering chili. Maria sat at a corner table, her Canon camera cradled in her hands, the worn leather strap looped around her wrist. She wasn't taking pictures now, just staring at the small LCD screen, replaying the same two frames over and over. A shadow fell over the table. She looked up to see a man in a rumpled tan jacket, a silver badge clipped to his belt. His face was set in a neutral, professional expression, but his eyes were sharp. "Señor," *she said, offering a small, polite smile.* *Hunter pulled out the chair opposite her and sat down, his gaze flicking from her face to the camera in her hands.* "Miss... it's Maria, right? Maria Vargas?" *She nodded, her smile faltering slightly.* "Sí. Yes." *He leaned forward, his voice low.* "I'm Detective Hunter, Alborn PD. I'm looking into the disappearance of a man named Elias Thorne. His shop, the clock repair place on the corner? It's been closed for two days. His neighbor said she saw him near this cafe the other evening, talking to someone. Your photo showed up on the street camera feed from that night. You were taking pictures." Maria’s brow furrowed as she listened, catching the gist of his words. "Desaparecido? A man... is missing?" *She translated the name with effort.* "Elias. The clock man," *she said, her voice suddenly tight.* "Sí. I saw him. That night." *Hunter's eyes narrowed.* "You did? What happened?" A wave of unease washed over Maria's face. She set the camera down on the worn wooden table as if it were suddenly too heavy. Her hands trembled slightly as she began to speak, her voice dropping to a hushed, rapid Spanish. "No estaba bien, señor. No estaba bien. Estaba hablando solo, caminando rápido. Tenía los ojos muy abiertos... asustado. Me miró, como si no me viera. Como si viera algo detrás de mí. Luego... luego entró al callejón junto a la tienda de la Sra. Higgins. Pero no salió. Yo esperé. Conté hasta cien. No salió." *(He wasn't right, sir. He wasn't right. He was talking to himself, walking fast. His eyes were wide... scared. He looked at me, like he didn't see me. Like he saw something behind me. Then... then he went into the alley next to Mrs. Higgins's shop. But he didn't come out. I waited. I counted to one hundred. He didn't come out.)* *Hunter stared at her, the rapid-fire Spanish washing over him like static. He caught the word "callejón" and the frantic motion of her hand pointing towards the street. His jaw tightened in frustration.* "Slow down, ma'am. I'm sorry, I don't—" *He stopped, running a hand through his short hair.* "You say he went into an alley? Did you see anyone with him? Another man? A car?" Maria saw the look on his face—the same look she got when a professor asked a question she couldn't answer. He wasn't understanding. Desperation filled her. She picked up her camera, fumbling to press a button. The screen flickered to life, showing the photo she'd been staring at. It was a shot of Elias Thorne, his face half-turned, his expression a mask of pure terror. His eyes were wide, staring at the entrance of the dark alley. But what made her blood run cold was what was behind him at the alley's mouth. A faint, shimmering distortion in the air, like the heat rising from the pavement, but it was near dusk and the air was cold. *She turned the camera to Hunter, her finger jabbing at the screen.* "Allí. Mire. Ahí." *(There. Look. There.)* *Hunter squinted at the screen.* The image was a mix of shadow and grain, but the man's fear was undeniable. He saw the terrified expression, and he saw the alley. But the shimmering heat distortion just looked like a photographic error to him, a smudge on the lens. "Is that him? Elias?" *he asked, pointing. He saw a small town, a missing man, and a frightened witness. He didn't see the anomaly. He didn't feel the wrongness that had settled over Alborn in the last two days.* "Yo no sé, señor. Yo no lo conozco. Pero él, él..." *She tried to explain, to tell him the fear she saw wasn't just on Elias's face, but that it seemed to radiate from the air around him. The more she tried to explain, the more frustrated she became. She switched to her broken English, her vowels flattened by stress.* "He... before... he look... very afraid. He be, like, here *afraid* of alley. Not of... no man. Something... not there." "I need you to try to be more specific, Maria. What did you see? Colors? A shape?" Hunter pressed, his voice taking on a harder edge. He was a man used to dealing in facts, and this conversation was dissolving into a frustrating fog. "It not a 'what'! It a 'nothing'!" *Maria cried, nearly dropping her camera. Her eyes were wide, pleading.* "A nothing that... it *moves*! The air... it has a... a *dente*. A bite! The light where he stand, it... in the photo, it *bites* the alley." Hunter stared at her, utterly lost. A bite in the light. A moving nothing. He let out a slow, controlled breath. The poor girl was clearly shaken, maybe a little unhinged by the incident. Or maybe she was just a tourist not used to small-town quiet and had let her imagination run wild. He had a stack of missing-person reports and theories of a runaway on his desk that felt far more grounded. Just then, his radio crackled to life. "Detective Hunter, you there? We got a status update from the county. They're holding a possible John Doe matching Thorne's description at the St. Cloud hospital. Requesting you to respond." *Hunter’s posture changed instantly. He stood up, all trace of his earlier frustration gone, replaced by a professional urgency. He clipped his radio back to his belt.* "I have to go, Maria. *He clicked his radio.*"I'm on my way, dispatch. Out." *He looked back at her.* "Stay in town, alright? Don't go wandering off. I may have more questions for you." *Maria looked up at him, her face still pale, her hand clutching her camera like a lifeline.* "Señor, por favor. You... usted... you must see. The alley. It is not... good. *Malo*. It is bad place." *Hunter gave her a curt, professional nod, his mind already on the hospital and the possibility of a positive ID.* "I'll check it out. You take care of yourself." He turned and strode out of the cafe, the door's bell jingling his departure. Maria watched him go, a cold knot of dread tightening in her stomach. She looked back down at the photo on her camera. The shimmering distortion seemed to pulse, even in the still image. She whispered to the empty table, her words half to herself, half to the absent detective who couldn't understand. "Yo sé lo que vi. Muy mal. No es un hombre, señor. No es un hombre." *(I know what I saw. Very bad. It is not a man, sir. It is not a man.)*
created_at
"2026-09-03T09:50:20.473Z"
9/3/2026, 9:50:20 AM
model
"deepseek/deepseek-v4-flash-20260731"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
567
moderation_latency
(null)
generation_time
67861
tokens_prompt
304
tokens_completion
2199
native_tokens_prompt
263
native_tokens_completion
2048
native_tokens_completion_images
(null)
native_tokens_reasoning
352
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
0.00034083
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788429020-DfNZH2cho9JNBeZls7Z8"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788429020-9DBq4n8o7Zwp01pCpU1s"
upstream_id
"chatcmpl-1f093bb5-e7fc-4ab0-b072-5245bea47b19"
provider_responses
0
endpoint_id
"562d5188-97d1-49f7-8881-d0e417e31ada"
is_byok
false
latency
46
model_permaslug
"deepseek/deepseek-v4-flash-20260731"
provider_name
"Makora"
status
429
1
endpoint_id
"f954a48e-1c90-433b-9348-4720f8030331"
id
"chatcmpl-1f093bb5-e7fc-4ab0-b072-5245bea47b19"
is_byok
false
latency
567
model_permaslug
"deepseek/deepseek-v4-flash-20260731"
provider_name
"OpenInference"
status
200
total_cost
0.00034083
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenInference"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
37.50%
Parse dialogue
n/a
dialogue
0
name
"Hunter"
dialogue
"Miss... it's Maria, right? Maria Vargas?"
detectedLang
"en"
heavyLang
"en"
scores
en
0.754601226993865
tl
0.6860282574568289
es
0.5525727069351231
lt
0.504950495049505
sq
0.504950495049505
no
0.5037220843672456
da
0.5
de
0.494949494949495
pt
0.4350282485875706
it
0.4269340974212034
ro
0.42363112391930835
ca
0.42028985507246375
fr
0.4186046511627907
eu
0.41690962099125367
et
0.41348973607038125
yo
0.4065281899109792
sv
0.38650306748466257
ms
0.37888198757763975
nl
0.37888198757763975
fi
0.3710691823899371
hr
0.2727272727272727
sl
0.2647058823529412
pl
0.250936329588015
sk
0.23954372623574144
cs
0.2277992277992278
lv
0.2277992277992278
reliable
true
passes
true
1
name
"Maria"
dialogue
"Sí. I saw him. That night."
detectedLang
"en"
heavyLang
"en"
scores
en
0.8140276301806588
tl
0.6823956442831215
yo
0.6384297520661157
ca
0.5752427184466019
no
0.5558375634517766
da
0.5489690721649484
hr
0.44267515923566875
sv
0.44089456869009586
sq
0.4336569579288026
pl
0.4186046511627907
it
0.396551724137931
cs
0.375
ms
0.36363636363636365
es
0.3613138686131387
ro
0.33962264150943394
de
0.29718875502008035
fr
0.29718875502008035
sk
0.29149797570850206
be
0.2857142857142857
tr
0.2857142857142857
hu
0.2768595041322314
az
0.2647058823529412
et
0.2647058823529412
is
0.2647058823529412
lt
0.2647058823529412
lv
0.2647058823529412
uk
0.2647058823529412
vi
0.2647058823529412
reliable
true
passes
false
2
name
"Hunter"
dialogue
"You did? What happened?"
detectedLang
"en"
heavyLang
"en"
scores
en
0.8383620689655172
tl
0.5426829268292682
cs
0.34210526315789475
reliable
true
passes
true
3
name
"Hunter"
dialogue
"callejón"
detectedLang
"is"
heavyLang
"es"
scores
is
0.609375
en
0.5856353591160222
es
0.5562130177514794
de
0.5098039215686274
reliable
true
passes
false
4
name
"Maria"
dialogue
"A nothing that... it moves! The air... it has a... a dente. A bite! The light where he stand, it... in the photo, it bites the alley."
detectedLang
"en"
heavyLang
"en"
scores
en
0.8035506402793946
tl
0.6384574183181574
fr
0.6156036446469249
de
0.5423728813559322
nl
0.5018450184501845
it
0.4866920152091255
no
0.458700882117081
da
0.4569589702333065
yo
0.45608380338436744
ca
0.4439868204283361
ro
0.432296047098402
es
0.42845046570702794
ms
0.40685413005272403
sq
0.40423654015887023
pt
0.39624329159212884
sl
0.3209255533199195
sv
0.3154158215010142
cs
0.3041237113402062
sk
0.2887249736564805
pl
0.28722280887011614
fi
0.24916573971078973
et
0.24496644295302009
az
0.24327354260089684
is
0.24242424242424243
lt
0.2398648648648649
hr
0.2232451093210587
hu
0.18772563176895307
lv
0.18772563176895307
eu
0.17982989064398544
tr
0.1758241758241758
hi
0.10239361702127658
vi
0.09395973154362416
reliable
true
passes
false
5
name
"Maria"
dialogue
"Stay in town, alright? Don't go wandering off. I may have more questions for you."
detectedLang
"en"
heavyLang
"en"
scores
en
0.8146186440677966
tl
0.6247319513938527
da
0.5843230403800476
no
0.5222929936305732
it
0.48326771653543305
nl
0.4675456389452333
yo
0.44678609062170704
fr
0.44561774023231254
de
0.43426724137931033
sv
0.4061085972850679
pl
0.37722419928825623
ca
0.34782608695652173
pt
0.3312101910828026
sl
0.3155149934810952
ro
0.3064729194187582
hr
0.29245283018867924
cs
0.2584745762711864
ms
0.24132947976878616
vi
0.2290748898678414
az
0.21875
sq
0.2033383915022762
es
0.1264559068219634
sk
0.12060301507537688
be
0.11764705882352942
tr
0.11764705882352942
hu
0.11317567567567569
et
0.10714285714285714
is
0.10714285714285714
lt
0.10714285714285714
lv
0.10714285714285714
uk
0.10714285714285714
reliable
true
passes
false
6
name
"Maria"
dialogue
"Señor, por favor. You... usted... you must see. The alley. It is not... good. Malo. It is bad place."
detectedLang
"en"
heavyLang
"en"
scores
en
0.7764054514480409
es
0.6090841399851079
tl
0.5535714285714286
fr
0.45482866043613707
nl
0.44561774023231254
da
0.43910256410256415
pt
0.4101123595505618
sq
0.4061085972850679
no
0.38307873090481787
yo
0.36594202898550726
et
0.32605905006418484
cs
0.2981283422459893
ca
0.29435483870967744
ro
0.29245283018867924
de
0.29149797570850206
sv
0.28765264586160105
it
0.24132947976878616
sl
0.2324561403508772
hu
0.22680412371134018
eu
0.20212765957446807
is
0.14910858995137766
hr
0.1435562805872757
sk
0.1435562805872757
lv
0.14075286415711946
ms
0.12935323383084577
tr
0.12935323383084577
pl
0.12790697674418602
fi
0.11764705882352942
reliable
true
passes
false
7
name
"Hunter"
dialogue
"I'll check it out. You take care of yourself."
detectedLang
"en"
heavyLang
"en"
scores
en
0.8214285714285714
tl
0.6470588235294117
yo
0.38650306748466257
ro
0.3839835728952772
nl
0.29577464788732394
no
0.23857868020304565
is
0.22879177377892032
lv
0.2227979274611399
sq
0.2227979274611399
de
0.21671018276762402
cs
0.20634920634920634
sl
0.20634920634920634
reliable
true
passes
true