NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-10-08-01M4DA6B5VXJ0ECFCBHS4VY16X
Run Details
Model:
Claude Haiku 5.5 (Adaptive)
Test:
Language Writing
Scenario:
Character dialogue (Italian) in a story
Duration:
13.6s
Cost:
$0.001274
Provider:
Claude Platform on AWS
Evaluation summary
Score
Evaluator
Details Link
100.00%
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="Isabella"> Isabella is a 30-year old college student on a study abroad program. Born in Italy, she is fluent in Italian 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 Italian, 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: Isabella 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 Italian, and Hunter looks confused. They talk for a while, until Hunter gets a call and has to leave. Return the dialogue between Isabella and Hunter as a voice sheet: Isabella: "..." Hunter: "..."
Model response
Messages
Meta
AI
**VOICE SHEET** **Scene:** Café on Main Street, Alborn, MN. Late afternoon. Isabella sits by the window, turning her camera over in her hands. Hunter enters, scanning the room, and stops at her table. --- **Hunter:** Excuse me. Are you Isabella Ferraro? The woman from the Hendricks orchard photos? **Isabella:** *(hesitant, looking up)* Mi scusi... non capisco. Sono... sono italiana. Parlo solo italiano. *[Excuse me... I don't understand. I'm... I'm Italian. I only speak Italian.]* **Hunter:** *(pulling out a chair, sitting anyway)* Okay. I'm Detective Hunter. I need to ask you about Tuesday night. The night Mark Dorsey disappeared. **Isabella:** *(voice trembling, gripping the camera tighter)* Non è possibile. Ho visto qualcosa nella nebbia, vicino ai meli. Due figure. Una di loro si è fermata a guardarmi. Ho avuto tanta paura. *[It's not possible. I saw something in the fog, near the apple trees. Two figures. One of them stopped and looked at me. I was so scared.]* **Hunter:** *(frowning, leaning forward)* Two figures. Okay. Did they say anything? Did you see their faces? **Isabella:** *(shaking her head, then pointing urgently at the camera)* Le foto. Guardi. Le ho scattate prima che la nebbia diventasse troppo fitta. Non ho capito cosa stessi fotografando finché non sono tornata a casa. *[The photos. Look. I took them before the fog got too thick. I didn't realize what I was photographing until I got home.]* **Hunter:** *(holding out a hand, palm up)* I'm going to need to see those. But I'm going to be honest with you, I have no idea what you just said. **Isabella:** *(exhaling, almost laughing despite herself)* Ah, certo. Lei non parla italiano. Va bene. *(She scrolls through the camera's screen, then turns it toward him, tapping a blurred image of two dark shapes between rows of trees.)* Questa. Qui. *[Ah, of course. You don't speak Italian. Fine. This one. Here.]* **Hunter:** *(squinting at the screen, then glancing back at her)* Okay, I see the shapes. You want me to look at this one? **Isabella:** *(nodding quickly, pointing to the edge of the frame)* Sì, sì. Il cancello. Il cancello era aperto. Qualcuno l'ha lasciato aperto. *[Yes, yes. The gate. The gate was open. Someone left it open.]* **Hunter:** *(tapping the screen, muttering)* The gate was open. Right. Can you send me these? Email, text, anything. **Isabella:** *(pausing, then pulling out her phone and holding it up with a helpless shrug)* Non capisco la sua domanda. Ma posso mandarle le foto, se mi dice come. *[I don't understand your question. But I can send you the photos, if you tell me how.]* **Hunter:** *(taking out his own phone, pointing at it, then at hers)* Okay. Let's try this. Here. *(He gestures at the screen, mimes tapping, and slides his phone across the table.)* **Isabella:** *(smiling faintly, tapping away)* Ecco. Grazie. Lei è molto paziente, detective. *[There. Thank you. You're very patient, detective.]* **Hunter:** *(not understanding, but nodding anyway)* Yeah, I'm working on it. *(His phone buzzes loudly. He glances at the screen, and his expression hardens.)* **Hunter:** *(standing abruptly)* I have to go. Listen to me. Stay in town, and don't go anywhere near the Hendricks orchard. Understand? *(He pauses, then points at her, then at the door, then makes a flat "stay" gesture with his hand.)* **Isabella:** *(rising slowly, confused, clutching the camera)* Mi dispiace, non ho capito. Ma... faccia attenzione, detective. Quelle figure... sembravano sapere che io ero lì. *[I'm sorry, I didn't understand. But... be careful, detective. Those figures... they seemed to know I was there.]* **Hunter:** *(already walking toward the door, phone pressed to his ear)* Yeah. Hunter. Where? *(He pauses at the threshold and looks back at her.)* Just stay put. *(The door swings shut behind him. Isabella stares at the screen of her camera, where the blurred figures still watch from the fog.)* ---
created_at
"2026-10-08T08:31:03.875Z"
10/8/2026, 8:31:03 AM
model
"anthropic/claude-haiku-5.5-20261007"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
13619
moderation_latency
(null)
generation_time
13619
tokens_prompt
308
tokens_completion
1006
native_tokens_prompt
424
native_tokens_completion
2463
native_tokens_completion_images
(null)
native_tokens_reasoning
838
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
"end_turn"
service_tier
"default"
usage
0.0012739
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1791448263-a0MNstUqVNQXYy92R00u"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1791448263-kcYMsHNve2ZudPIbx4Cq"
upstream_id
"msg_011CfpTWvZ286S7KmV6XEoHJ"
provider_responses
0
endpoint_id
"b6cdf493-3a61-441f-a134-e7944fddb980"
id
"msg_011CfpTWvZ286S7KmV6XEoHJ"
is_byok
false
latency
785
model_permaslug
"anthropic/claude-haiku-5.5-20261007"
provider_name
"Claude Platform on AWS"
status
200
total_cost
0.0012739
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Claude Platform on AWS"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
100.00%
Parse dialogue
n/a
dialogue
0
name
"Hunter"
dialogue
"Excuse me. Are you Isabella Ferraro? The woman from the Hendricks orchard photos?"
detectedLang
"en"
heavyLang
"en"
scores
en
0.7094631986718318
tl
0.5454545454545454
fr
0.4883040935672514
nl
0.4675456389452333
sq
0.46428571428571425
no
0.44678609062170704
es
0.39306358381502887
it
0.37648456057007124
fi
0.3526510480887793
ca
0.33375634517766495
ms
0.33035714285714285
pt
0.33035714285714285
de
0.31906614785992216
ro
0.3027888446215139
da
0.26264044943820225
sv
0.25426136363636365
et
0.2532005689900427
cs
0.22106824925816024
yo
0.16930379746835442
sl
0.1435562805872757
is
0.14075286415711946
lt
0.13223140495867766
hi
0.12790697674418602
hr
0.1191275167785235
lv
0.1191275167785235
reliable
true
passes
true
1
name
"Isabella"
dialogue
"Mi scusi... non capisco. Sono... sono italiana. Parlo solo italiano."
detectedLang
"it"
heavyLang
"it"
scores
it
0.8074807480748075
es
0.7168284789644014
pt
0.6111111111111112
fr
0.5308310991957105
ro
0.5295698924731184
ca
0.5218579234972678
en
0.515235457063712
pl
0.4615384615384615
sq
0.45226917057902977
yo
0.4006849315068493
sl
0.39130434782608703
sk
0.38917975567190227
tl
0.38917975567190227
cs
0.36823104693140796
hr
0.36363636363636365
tr
0.35185185185185186
hu
0.3027888446215139
az
0.2693110647181628
de
0.2647058823529412
lt
0.2647058823529412
nl
0.2647058823529412
ms
0.2584745762711864
no
0.2584745762711864
da
0.25213675213675213
sv
0.25213675213675213
eu
0.19908466819221968
et
0.15254237288135594
fi
0.14425427872860636
lv
0.14425427872860636
reliable
true
passes
true
2
name
"Hunter"
dialogue
"Okay. I'm Detective Hunter. I need to ask you about Tuesday night. The night Mark Dorsey disappeared."
detectedLang
"en"
heavyLang
"en"
scores
en
0.7996300863131935
tl
0.6364653243847874
no
0.47197400487408614
da
0.463253509496284
fr
0.4318181818181818
it
0.4025735294117647
sv
0.39759036144578314
nl
0.38679245283018865
yo
0.3838862559241706
ro
0.38154138915318747
cs
0.3725868725868726
ca
0.3670886075949367
sq
0.358974358974359
hr
0.33127572016460904
pl
0.2864983534577387
sl
0.2753623188405797
de
0.27455357142857145
ms
0.2655367231638418
pt
0.2638731596828992
sk
0.2388758782201405
et
0.23708920187793428
lv
0.2215568862275449
es
0.214975845410628
lt
0.20440636474908203
eu
0.1761723700887199
is
0.1761723700887199
tr
0.16559691912708602
be
0.09722222222222222
hu
0.09344490934449094
az
0.08835904628330996
uk
0.08835904628330996
vi
0.08835904628330996
reliable
true
passes
true
3
name
"Isabella"
dialogue
"Non è possibile. Ho visto qualcosa nella nebbia, vicino ai meli. Due figure. Una di loro si è fermata a guardarmi. Ho avuto tanta paura."
detectedLang
"it"
heavyLang
"it"
scores
it
0.8144168782965423
ca
0.5929734361610968
es
0.5299356754082137
ro
0.5228528377699648
fr
0.5175215845606908
en
0.5123203285420944
pt
0.49788583509513745
tl
0.45621064682312534
sq
0.3994943109987357
yo
0.35020519835841313
sl
0.3448275862068966
sk
0.31457431457431456
no
0.3035190615835778
de
0.2846385542168675
cs
0.2731446059678654
fi
0.27035330261136714
et
0.258391881342701
lt
0.24302788844621515
sv
0.24242424242424243
ms
0.24060751398880897
lv
0.23263327948303714
vi
0.223221586263287
hr
0.20767306088407003
az
0.19898819561551434
is
0.19011082693947143
da
0.1831470335339639
eu
0.14182475158084915
hu
0.1410488245931284
pl
0.0995260663507109
tr
0.07497565725413825
nl
0.06862745098039216
reliable
true
passes
true
4
name
"Hunter"
dialogue
"Two figures. Okay. Did they say anything? Did you see their faces?"
detectedLang
"en"
heavyLang
"en"
scores
en
0.8220640569395018
tl
0.6039603960396039
fr
0.4482758620689655
ca
0.43741209563994377
et
0.40119760479041916
ro
0.3442622950819672
es
0.33884297520661155
sq
0.30675909878682844
ms
0.26062846580406657
cs
0.25512104283054005
it
0.2537313432835821
da
0.24812030075187969
de
0.24812030075187969
no
0.24242424242424243
pt
0.24242424242424243
lv
0.17184265010351968
hu
0.13606911447084233
pl
0.13606911447084233
sk
0.13606911447084233
fi
0.12854030501089325
hr
0.12854030501089325
nl
0.12854030501089325
sl
0.12854030501089325
reliable
true
passes
true
5
name
"Isabella"
dialogue
"Le foto. Guardi. Le ho scattate prima che la nebbia diventasse troppo fitta. Non ho capito cosa stessi fotografando finché non sono tornata a casa."
detectedLang
"it"
heavyLang
"it"
scores
it
0.8013383521539105
es
0.6252465483234714
ca
0.6109746109746109
ro
0.5838808585194919
pt
0.5539906103286385
fr
0.5465393794749402
tl
0.4611457742484402
hr
0.4574528840662478
en
0.44476914085330216
no
0.41896024464831805
sl
0.40736119775421087
et
0.40624999999999994
sk
0.40213971050975456
cs
0.39911448450347875
de
0.38985228002569045
sq
0.3679308050565535
da
0.36241610738255037
nl
0.35767410412440837
ms
0.35153583617747436
sv
0.31605471562275017
pl
0.2813918305597579
lt
0.26810477657935283
yo
0.25954793452844893
eu
0.24662965900079303
tr
0.21810699588477367
lv
0.21161825726141079
hu
0.21096345514950168
fi
0.18032786885245902
az
0.17606244579358196
is
0.16373239436619716
vi
0.1394927536231884
reliable
true
passes
true
6
name
"Hunter"
dialogue
"I'm going to need to see those. But I'm going to be honest with you, I have no idea what you just said."
detectedLang
"en"
heavyLang
"en"
scores
en
0.8329621380846325
et
0.5584524810765349
tl
0.4760479041916168
da
0.4712990936555891
no
0.4496855345911949
ca
0.4485294117647059
hr
0.4040862656072645
pl
0.3902439024390244
yo
0.38882421420256114
lv
0.37574316290130794
cs
0.37425506555423127
sk
0.3597560975609756
lt
0.34701492537313433
sv
0.3137254901960784
sq
0.308300395256917
es
0.29906542056074764
fr
0.27586206896551724
it
0.27083333333333337
hu
0.2595204513399154
tr
0.23133235724743775
ms
0.22794117647058826
eu
0.22680412371134018
pt
0.2081447963800905
sl
0.1873065015479876
ro
0.14634146341463414
be
0.11764705882352942
az
0.10714285714285714
is
0.10714285714285714
uk
0.10714285714285714
vi
0.10714285714285714
reliable
true
passes
true
7
name
"Isabella"
dialogue
"Ah, certo. Lei non parla italiano. Va bene. Questa. Qui."
detectedLang
"it"
heavyLang
"it"
scores
it
0.8
pt
0.6716287215411559
ca
0.6666666666666666
es
0.6612466124661246
fr
0.6518105849582173
ro
0.5932754880694143
en
0.5192307692307693
tl
0.48700410396716826
az
0.3811881188118812
tr
0.3697478991596639
sq
0.3676222596964587
sl
0.3567753001715266
hr
0.34782608695652173
cs
0.3409490333919156
sv
0.33510638297872336
ms
0.32795698924731176
no
0.32795698924731176
fi
0.30167597765363136
is
0.2951127819548872
sk
0.27884615384615385
pl
0.27466150870406186
lv
0.25742574257425743
de
0.251497005988024
lt
0.251497005988024
nl
0.251497005988024
da
0.23935091277890466
eu
0.1883116883116883
yo
0.15730337078651685
hu
0.1515837104072398
et
0.14383561643835618
reliable
true
passes
true
8
name
"Hunter"
dialogue
"Okay, I see the shapes. You want me to look at this one?"
detectedLang
"en"
heavyLang
"en"
scores
en
0.8210633946830265
tl
0.6974935177182369
no
0.6292372881355932
da
0.5862884160756502
sq
0.5070422535211268
pl
0.49856733524355296
hr
0.4942196531791907
yo
0.48604992657856094
et
0.48377581120943947
cs
0.4728915662650602
nl
0.45987654320987653
sv
0.38162544169611307
sl
0.3738819320214669
sk
0.36823104693140796
it
0.35779816513761464
lv
0.34579439252336447
fr
0.325626204238921
lt
0.3230174081237911
ms
0.31640625
ca
0.2827868852459016
es
0.22566371681415928
ro
0.20454545454545453
pt
0.2009132420091324
fi
0.1764705882352941
be
0.16666666666666666
tr
0.16666666666666666
hu
0.1606714628297362
az
0.15254237288135594
is
0.15254237288135594
uk
0.15254237288135594
vi
0.15254237288135594
reliable
true
passes
true
9
name
"Isabella"
dialogue
"Sì, sì. Il cancello. Il cancello era aperto. Qualcuno l'ha lasciato aperto."
detectedLang
"it"
heavyLang
"it"
scores
it
0.8125937031484258
es
0.6231155778894472
pt
0.5906113537117904
ca
0.5471014492753623
fr
0.503968253968254
ro
0.4747899159663866
yo
0.3951612903225807
en
0.3872549019607844
tl
0.28977272727272724
az
0.23469387755102045
fi
0.19871794871794873
eu
0.19006479481641467
de
0.1830065359477124
tr
0.1830065359477124
sv
0.1648106904231626
lt
0.1610738255033557
lv
0.1610738255033557
ms
0.1610738255033557
reliable
true
passes
true
10
name
"Hunter"
dialogue
"The gate was open. Right. Can you send me these? Email, text, anything."
detectedLang
"en"
heavyLang
"en"
scores
en
0.817444219066937
tl
0.5871559633027523
sq
0.5273109243697479
no
0.4943820224719101
nl
0.46619217081850534
ms
0.4201030927835051
yo
0.3732590529247911
fr
0.35064935064935066
cs
0.3401759530791789
de
0.3001555209953344
it
0.2857142857142857
ca
0.256198347107438
sv
0.2537313432835821
tr
0.2537313432835821
da
0.2524916943521595
es
0.1847826086956522
eu
0.18032786885245902
et
0.17431192660550457
pt
0.16356877323420074
sl
0.16356877323420074
az
0.14611005692599618
fi
0.14285714285714285
sk
0.14285714285714285
reliable
true
passes
true
11
name
"Isabella"
dialogue
"Non capisco la sua domanda. Ma posso mandarle le foto, se mi dice come."
detectedLang
"it"
heavyLang
"it"
scores
it
0.8033940397350994
fr
0.698029243483789
es
0.6953175112251443
ro
0.6607142857142858
pt
0.6444610778443113
sl
0.6131921824104235
ca
0.5994940978077572
yo
0.5685740236148955
en
0.5527306967984934
hr
0.5057232049947971
tl
0.4930629669156884
sq
0.4398584905660377
no
0.4358669833729216
sk
0.4290865384615385
cs
0.41430332922318125
et
0.3716931216931217
sv
0.36918990703851257
hu
0.3581081081081081
pl
0.343015214384509
da
0.33286516853932585
lv
0.29210134128166915
ms
0.29210134128166915
vi
0.2543171114599686
eu
0.24841772151898736
tr
0.23633440514469453
lt
0.21746293245469525
fi
0.17677642980935876
az
0.16225749559082892
is
0.14721723518850988
de
0.1171003717472119
nl
0.1171003717472119
reliable
true
passes
true
12
name
"Hunter"
dialogue
"Okay. Let's try this. Here."
detectedLang
"en"
heavyLang
"en"
scores
en
0.7302158273381295
tl
0.5774647887323944
nl
0.39999999999999997
sq
0.39999999999999997
reliable
true
passes
true
13
name
"Isabella"
dialogue
"Ecco. Grazie. Lei è molto paziente, detective."
detectedLang
"it"
heavyLang
"it"
scores
it
0.8057381948595339
ro
0.6516613076098606
pt
0.5922208281053952
ca
0.5849297573435505
es
0.5504840940525588
en
0.5276162790697675
fr
0.5170876671619613
tl
0.49768160741885625
sq
0.4217081850533808
no
0.411231884057971
de
0.40909090909090906
da
0.40476190476190477
nl
0.3486973947895792
lv
0.3200836820083682
hr
0.30851063829787234
yo
0.29805615550755943
cs
0.2872807017543859
sk
0.2872807017543859
sv
0.2793791574279379
pl
0.2243436754176611
eu
0.1935483870967742
sl
0.1709183673469388
lt
0.15364583333333331
reliable
true
passes
true
14
name
"Hunter"
dialogue
"Yeah, I'm working on it."
detectedLang
"en"
heavyLang
"en"
scores
en
0.8127340823970037
tl
0.7185741088180112
fr
0.5454545454545454
et
0.506578947368421
yo
0.5016611295681063
fi
0.5
nl
0.40476190476190477
ca
0.39759036144578314
hr
0.3697478991596639
lv
0.3644067796610169
no
0.3644067796610169
sq
0.3644067796610169
hi
0.358974358974359
pl
0.358974358974359
de
0.34782608695652173
cs
0.3333333333333333
sk
0.3333333333333333
is
0.3212669683257918
pt
0.3212669683257918
sv
0.3212669683257918
reliable
true
passes
true
15
name
"Hunter"
dialogue
"I have to go. Listen to me. Stay in town, and don't go anywhere near the Hendricks orchard. Understand?"
detectedLang
"en"
heavyLang
"en"
scores
en
0.7909295593438405
no
0.6247113163972287
da
0.6176470588235294
tl
0.5729303547963206
de
0.5451364590622813
nl
0.5419309372797745
sv
0.5181616011860637
pl
0.4791666666666667
yo
0.47791164658634533
sq
0.4352736750651607
sl
0.39365671641791045
it
0.38154138915318747
cs
0.37076476282671833
ro
0.36211972522080466
sk
0.3512974051896207
ms
0.3394308943089431
hr
0.336734693877551
fr
0.3207941483803553
is
0.3107104984093319
lv
0.2942453854505972
fi
0.28806133625410735
lt
0.20440636474908203
es
0.20245398773006135
et
0.19554455445544552
pt
0.18851435705368288
az
0.18444165621079048
hu
0.18444165621079048
ca
0.1751269035532995
tr
0.17407878017789072
eu
0.15254237288135594
be
0.09722222222222222
uk
0.08835904628330996
vi
0.08835904628330996
reliable
true
passes
true
16
name
"Isabella"
dialogue
"Mi dispiace, non ho capito. Ma... faccia attenzione, detective. Quelle figure... sembravano sapere che io ero lì."
detectedLang
"it"
heavyLang
"it"
scores
it
0.7977891614990563
fr
0.6181262729124236
ro
0.5247148288973384
en
0.5210727969348659
ca
0.48700410396716826
es
0.410840534171249
de
0.39271255060728744
tl
0.38825448613376834
sk
0.3718592964824121
no
0.3686868686868687
pt
0.36278674596431604
sq
0.3466898954703833
yo
0.32795698924731176
cs
0.3049119555143651
sl
0.2951127819548872
hr
0.2857142857142857
pl
0.2711370262390671
ms
0.2596248766041461
da
0.2537313432835821
et
0.22760041194644695
hu
0.18211559432933477
fi
0.13693901035673187
eu
0.1039426523297491
nl
0.0985576923076923
vi
0.09420289855072464
lv
0.09200968523002422
tr
0.08759124087591241
sv
0.08200734394124846
lt
0.07975460122699386
reliable
true
passes
true
17
name
"Hunter"
dialogue
"Yeah. Hunter. Where? Just stay put."
detectedLang
"en"
heavyLang
"en"
scores
en
0.798183652875883
sv
0.46236559139784944
de
0.453551912568306
sq
0.4065281899109792
ca
0.40476190476190477
hr
0.38461538461538464
et
0.33774834437086093
da
0.2907801418439716
no
0.2805755395683453
fr
0.2647058823529412
tl
0.250936329588015
it
0.23954372623574144
nl
0.23954372623574144
pt
0.23954372623574144
ms
0.2277992277992278
ro
0.2277992277992278
sl
0.2277992277992278
tr
0.2277992277992278
reliable
true
passes
true