NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-20-01KHY7KXDW0KXSDAY3QE1X8A8J
Run Details
Model:
Ministral 3 3B
Test:
Language Writing
Scenario:
Character dialogue (Hindi) in a story
Duration:
2.5s
Cost:
$0.000094
Provider:
Mistral
Evaluation summary
Score
Evaluator
Details Link
66.67%
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="Selvi"> Selvi is a 30-year old college student on a study abroad program. Born in India, she is fluent in Hindi 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 Hindi, 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: Selvi 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 Hindi, and Hunter looks confused. They talk for a while, until Hunter gets a call and has to leave. Return the dialogue between Selvi and Hunter as a voice sheet: Selvi: "..." Hunter: "..."
Model response
Messages
Meta
AI
**Title: *Whispers in the Apple Orchard*** --- **Scene: *The Morning Sunlight Through the Café Window*** *(Selvi sits at a small wooden table in *The Orchard Café*, her camera resting on the table between them. The scent of freshly brewed coffee and cinnamon fills the air. She adjusts the lens of her camera, her fingers moving with practiced ease. Across from her, Hunter leans against the counter, arms crossed, his sharp eyes scanning the room. He doesn’t look away from her as he steps forward.)* --- **Hunter:** *(approaching, voice low and measured)* "Selvi. You’re the one who took those photos of the old mill last week, right?" *(Selvi doesn’t look up. She keeps her camera steady, her expression unreadable. She speaks in Hindi, her voice barely above a whisper.)* **Selvi:** *"Aapko kya mila?"* *(What did you find?)* *(Hunter blinks. He’s never heard her speak before. The words roll off his tongue in English, but the tone—*shaken*, *hesitant*—hangs between them like a warning.)* **Hunter:** *(raising an eyebrow, voice tight)* "What did I find?" *(Selvi exhales, her shoulders tense. She glances at her camera again, then back at him. Her voice drops, urgent.)* **Selvi:** *"Aapko *dushman* mila."* *(You found a stranger.)* *(Hunter’s jaw tightens. He doesn’t know Hindi, but the way she says it—*dangerous*, *unsettled*—makes his instincts flare. He steps closer, his voice dropping to a growl.)* **Hunter:** "Who?" *(Selvi hesitates, then pulls out her phone. She shows him a photo—a man in a faded jacket, standing near the old mill. His face is blurred, but the way he lingers, *watching*, is unmistakable.)* **Selvi:** *"Aapko *uske* *pahuch* gaya."* *(You were followed.)* *(Hunter’s grip on his coffee cup tightens. He doesn’t understand the words, but the *meaning*—the way she clenches her jaw—tells him something’s off. He studies the photo, then looks up at her.)* **Hunter:** "You’re sure this is who?" *(Selvi nods, her voice barely audible.)* **Selvi:** *"Yes. *Aapko* *khub* *dushman* mila."* *(You found a very dangerous man.)* *(Hunter exhales sharply, rubbing his temple. His phone buzzes—a text from his partner. He glances at it, then back at Selvi.)* **Hunter:** "I gotta take this. But I’ll be back. You stay here." *(Selvi doesn’t move. She watches him leave, her camera still in hand. The café door swings shut behind him, and the silence that follows is heavy with unspoken questions.)* --- **End Scene.** *(Next: Hunter races to the mill, only to find something far worse than he expected.)*
created_at
"2026-02-20T19:14:53.513Z"
2/20/2026, 7:14:53 PM
model
"mistralai/ministral-3b-2512"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
275
moderation_latency
(null)
generation_time
1988
tokens_prompt
315
tokens_completion
694
native_tokens_prompt
274
native_tokens_completion
669
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
0.0000943
router
(null)
provider_responses
0
endpoint_id
"b9f6828f-aef9-498a-91f5-6816fbf72420"
id
"5f6aba596fbc4a4181b8efdf05efb142"
is_byok
false
latency
275
model_permaslug
"mistralai/ministral-3b-2512"
provider_name
"Mistral"
status
200
api_type
"completions"
id
"gen-1771614893-v88drLygzosJ0gwD65Ap"
upstream_id
"5f6aba596fbc4a4181b8efdf05efb142"
total_cost
0.0000943
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Mistral"
Evaluation details
Result
Evaluator
Details
Meta Data
66.67%
Parse dialogue
n/a
dialogue
0
name
"Hunter"
dialogue
"Selvi. You’re the one who took those photos of the old mill last week, right?"
detectedLang
"en"
heavyLang
"en"
scores
en
0.8099581908019764
tl
0.6299037749814952
fr
0.5112414467253177
nl
0.4807892004153686
no
0.41588785046728977
yo
0.37027707808564225
da
0.34554973821989526
de
0.3131868131868132
sq
0.2625368731563422
ca
0.22839506172839508
fi
0.18032786885245902
is
0.15110356536502548
hr
0.1482112436115843
pl
0.1482112436115843
es
0.13793103448275862
lt
0.13793103448275862
pt
0.13793103448275862
ms
0.13494809688581316
sv
0.13494809688581316
et
0.1334488734835355
hi
0.1334488734835355
lv
0.1243432574430823
reliable
true
passes
true
1
name
"Selvi"
dialogue
"Aapko kya mila?"
detectedLang
"it"
heavyLang
"fi"
scores
it
0.5555555555555556
reliable
false
passes
false
2
name
"Hunter"
dialogue
"What did I find?"
detectedLang
"en"
heavyLang
"en"
scores
en
0.8355263157894737
da
0.7050147492625368
no
0.5918367346938775
hr
0.5815899581589958
ca
0.5798319327731093
sv
0.5798319327731093
sq
0.5726495726495726
pl
0.5575221238938053
it
0.5348837209302325
cs
0.5121951219512195
ro
0.47368421052631576
yo
0.46524064171123
tl
0.42857142857142855
sk
0.4186046511627907
be
0.4117647058823529
tr
0.4117647058823529
hu
0.40119760479041916
az
0.38650306748466257
et
0.38650306748466257
is
0.38650306748466257
lt
0.38650306748466257
lv
0.38650306748466257
uk
0.38650306748466257
vi
0.38650306748466257
reliable
true
passes
true
3
name
"Selvi"
dialogue
"Aapko mila."
detectedLang
"it"
heavyLang
"fi"
scores
it
0.625
reliable
true
passes
false
4
name
"Hunter"
dialogue
"Who?"
detectedLang
"en"
heavyLang
"en"
scores
en
0.8626373626373627
reliable
false
passes
true
5
name
"Selvi"
dialogue
"Aapko gaya."
detectedLang
"tl"
heavyLang
"id"
scores
tl
0.5989304812834225
ms
0.5714285714285714
reliable
true
passes
false
6
name
"Hunter"
dialogue
"You’re sure this is who?"
detectedLang
"en"
heavyLang
"en"
scores
en
0.8429084380610413
tl
0.5333333333333334
nl
0.4760479041916168
hu
0.46808510638297873
yo
0.4067796610169492
is
0.3445692883895131
reliable
true
passes
true
7
name
"Selvi"
dialogue
"Yes. mila."
detectedLang
"it"
heavyLang
"ber"
scores
it
0.7142857142857143
en
0.6774193548387096
reliable
false
passes
true
8
name
"Hunter"
dialogue
"I gotta take this. But I’ll be back. You stay here."
detectedLang
"en"
heavyLang
"en"
scores
en
0.8101635514018692
tl
0.6520342612419701
it
0.4905956112852664
hu
0.473257698541329
sq
0.4186046511627907
sv
0.3981481481481482
yo
0.37739463601532564
cs
0.3602362204724409
lt
0.35000000000000003
no
0.30851063829787234
da
0.2995689655172413
hr
0.2995689655172413
ca
0.29805615550755943
pl
0.2793791574279379
fi
0.2562929061784897
ro
0.21686746987951808
fr
0.21307506053268765
sl
0.1935483870967742
sk
0.181360201511335
be
0.17721518987341772
tr
0.17721518987341772
az
0.16237113402061853
et
0.16237113402061853
is
0.16237113402061853
lv
0.16237113402061853
uk
0.16237113402061853
vi
0.16237113402061853
reliable
true
passes
true