NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-03-25-01KMJ85QNPBWKKNPVD45GM95VK
Run Details
Model:
Qwen 3 32B
Test:
Language Writing
Scenario:
Character dialogue (Hindi) in a story
Duration:
36.2s
Cost:
$0.000355
Provider:
Nebius
Evaluation summary
Score
Evaluator
Details Link
62.50%
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="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
**Scene: Cozy Corner Café, Alborn, MN** *Selvi sits by the window, her camera resting on the wooden table. The autumn sunlight casts a warm glow over her hands as she frowns at a photo on her phone. The café hums with the low murmur of conversation and the clink of coffee mugs. Hunter approaches, his boots creaking against the floorboards, and leans against the table.* **Hunter:** (gruff, but not unkind) “Ma’am, I need to ask you a few questions about the… *incident* at the old orchard last night.” *Selvi’s eyes widen. She opens her mouth, then closes it, glancing at her camera as if for answers.* **Selvi:** (rapid Hindi, voice trembling) “Woh log… unko kyun chalaya? Maine… maine dekha tha, but—” *Her hands flutter, pointing at her phone’s screen. Hunter stares blankly.* **Hunter:** (rubbing the back of his neck) “Sorry. I don’t—do you speak English?” *Selvi shakes her head violently, her braid slipping over her shoulder. She taps her own temples, then points at the photo: a gnarled apple tree, its branches twisted like skeletal fingers.* **Selvi:** (Hindi, faster now) “Woh pani… ganda hai. Kuchh hua… kuchh *bhoot*—” **Hunter:** (interrupting, leaning closer) “Bhoot? You saw something at the orchard? A person? Animal?” *She clutches her camera, her knuckles pale.* **Selvi:** (Hindi, near a whisper) “Nahin… kuchh aur. Kali aankhen. Kali, *desh* ke…” *(“No… something else. Black eyes. Black, of the…”)* *Hunter’s jaw tightens. His radio crackles in his pocket. He pulls it out, face hardening.* **Hunter:** (into radio) “Alborn PD, Detective Hunter. Copy that.” *His gaze flicks back to Selvi, apologetic but firm. She’s already turning away, her camera shielding her face.* **Hunter:** (softly, in case she understands words) “We’ll figure this out.” *He walks away as the radio’s static swallows his next breath. Selvi stares at the photo, the tree’s shadow now looking like a figure with black eyes.* --- ** END SCENE **
created_at
"2026-03-25T10:20:49.722Z"
3/25/2026, 10:20:49 AM
model
"qwen/qwen3-32b-04-28"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
585
moderation_latency
(null)
generation_time
36183
tokens_prompt
339
tokens_completion
1190
native_tokens_prompt
280
native_tokens_completion
1091
native_tokens_completion_images
(null)
native_tokens_reasoning
642
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
0.0003553
router
(null)
provider_responses
0
endpoint_id
"aaa39b6c-e488-491a-91e4-e5fb66f86601"
id
"cmpl-9067249f43d77fc4"
is_byok
false
latency
578
model_permaslug
"qwen/qwen3-32b-04-28"
provider_name
"Nebius"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.0; linux; x64))"
http_referer
(null)
api_type
"completions"
id
"gen-1774434049-WuWdJewSmpQj9sdfoE41"
upstream_id
"cmpl-9067249f43d77fc4"
total_cost
0.0003553
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Nebius"
Evaluation details
Result
Evaluator
Details
Meta Data
62.50%
Parse dialogue
n/a
dialogue
0
name
"Hunter"
dialogue
"“Ma’am, I need to ask you a few questions about the… incident at the old orchard last night.”"
detectedLang
"en"
heavyLang
"en"
scores
en
0.8084584737971193
tl
0.6786632390745501
fr
0.5920365535248042
no
0.5756958587915818
da
0.5500359971202303
ca
0.5023885350318471
it
0.4955609362389023
nl
0.49186991869918695
cs
0.47786131996658315
yo
0.4560487380330722
et
0.42974452554744524
sk
0.42555147058823534
ro
0.4196843082636954
pt
0.41479400749063666
pl
0.410933081998115
hr
0.40532825880114176
sq
0.3914313534566699
es
0.35765673175745116
sv
0.35367114788004134
sl
0.35233160621761656
de
0.3509865005192108
lt
0.33011789924973206
hu
0.31091510474090406
az
0.3078626799557032
lv
0.28977272727272724
tr
0.19041450777202074
ms
0.18407310704960833
is
0.18086500655307994
vi
0.17546174142480211
be
0.10071942446043164
eu
0.10071942446043164
fi
0.09156976744186046
uk
0.09156976744186046
reliable
true
passes
true
1
name
"Selvi"
dialogue
"“Woh log… unko kyun chalaya? Maine… maine dekha tha, but—”"
detectedLang
"en"
heavyLang
"de"
scores
en
0.5810055865921787
fr
0.5689655172413793
tl
0.46428571428571425
et
0.35205183585313177
fi
0.35205183585313177
de
0.3288590604026846
ro
0.3288590604026846
sq
0.2805755395683453
ms
0.25925925925925924
az
0.25
lv
0.22680412371134018
reliable
false
passes
false
2
name
"Hunter"
dialogue
"“Sorry. I don’t—do you speak English?”"
detectedLang
"en"
heavyLang
"en"
scores
en
0.8055162659123054
tl
0.6469833119383827
hr
0.5462046204620462
cs
0.5346869712351946
pl
0.494485294117647
pt
0.4821092278719397
sq
0.47718631178707227
no
0.45219123505976094
it
0.4387755102040817
da
0.434156378600823
sv
0.43298969072164945
sk
0.4258872651356994
vi
0.38752783964365256
yo
0.375
ro
0.3707093821510297
ca
0.3341404358353511
sl
0.3037974683544304
de
0.207492795389049
ms
0.207492795389049
be
0.20289855072463767
tr
0.20289855072463767
hu
0.19590643274853803
az
0.1863905325443787
et
0.1863905325443787
is
0.1863905325443787
lt
0.1863905325443787
lv
0.1863905325443787
uk
0.1863905325443787
reliable
true
passes
true
3
name
"Selvi"
dialogue
"“Woh pani… ganda hai. Kuchh hua… kuchh bhoot—”"
detectedLang
"ms"
heavyLang
"de"
scores
ms
0.4747899159663866
sk
0.46120689655172414
it
0.4004796163069544
pl
0.39759036144578314
tl
0.3902439024390244
vi
0.3857493857493857
lv
0.37810945273631835
et
0.36061381074168797
ro
0.358974358974359
nl
0.29577464788732394
az
0.2690058479532163
is
0.24698795180722888
ca
0.2378048780487805
cs
0.23547400611620795
tr
0.23076923076923078
eu
0.21875
es
0.21135646687697163
lt
0.21135646687697163
pt
0.21135646687697163
no
0.20127795527156547
en
0.19093851132686085
hr
0.19093851132686085
sq
0.19093851132686085
sv
0.19093851132686085
reliable
false
passes
false
4
name
"Hunter"
dialogue
"“Bhoot? You saw something at the orchard? A person? Animal?”"
detectedLang
"en"
heavyLang
"en"
scores
en
0.7821350762527233
tl
0.6726677577741408
no
0.5798319327731093
pt
0.5614035087719298
fr
0.5604395604395604
da
0.5418098510882016
nl
0.5055624227441285
ms
0.5043370508054523
es
0.48586118251928023
de
0.47643979057591623
sv
0.4350282485875706
ca
0.43342776203966005
sq
0.4269340974212034
ro
0.42363112391930835
it
0.4152046783625731
cs
0.4020926756352765
hr
0.37694704049844235
yo
0.3610223642172524
hu
0.35379644588045234
is
0.34318555008210183
az
0.33884297520661155
sk
0.32088285229202035
pl
0.2907801418439716
lt
0.2779783393501805
lv
0.26605504587155965
tr
0.2647058823529412
et
0.23371647509578544
sl
0.1836734693877551
eu
0.14893617021276595
vi
0.14893617021276595
fi
0.13606911447084233
reliable
true
passes
true
5
name
"Selvi"
dialogue
"“Nahin… kuchh aur. Kali aankhen. Kali, desh ke…”"
detectedLang
"ms"
heavyLang
"tl"
scores
ms
0.5888157894736842
tl
0.4802494802494803
cs
0.4186046511627907
eu
0.37027707808564225
nl
0.31880108991825606
ro
0.29577464788732394
yo
0.26035502958579887
pl
0.23547400611620795
sk
0.23547400611620795
reliable
true
passes
false
6
name
"Hunter"
dialogue
"“Alborn PD, Detective Hunter. Copy that.”"
detectedLang
"en"
heavyLang
"en"
scores
en
0.7270742358078602
ro
0.62178517397882
tl
0.6124031007751938
it
0.5819397993311036
fr
0.5145631067961165
ca
0.4887525562372188
pt
0.48240165631469983
es
0.41588785046728977
de
0.4075829383886255
da
0.39903846153846156
no
0.3872549019607844
sq
0.37810945273631835
nl
0.3670886075949367
sv
0.3540051679586563
sl
0.33510638297872336
cs
0.22360248447204967
hr
0.21135646687697163
ms
0.19093851132686085
tr
0.19093851132686085
reliable
true
passes
true
7
name
"Hunter"
dialogue
"“We’ll figure this out.”"
detectedLang
"en"
heavyLang
"en"
scores
en
0.8235294117647058
tl
0.5714285714285714
fr
0.494949494949495
ro
0.494949494949495
ca
0.33920704845814975
sq
0.33920704845814975
da
0.3150684931506849
de
0.3150684931506849
es
0.3150684931506849
it
0.3150684931506849
no
0.3150684931506849
pt
0.3150684931506849
reliable
true
passes
true