NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-03-06-01KK1GHZG7QMX9P8C9MRXY6T97
Run Details
Model:
GPT-5.4
Test:
Dialogue tags
Scenario:
Write 200 words with 90% dialogue
Duration:
11.1s
Cost:
$0.005100
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
51.93%
Matches word count
(details)
36.79%
Dialogue to Total Word Ratio
(details)
44.3567%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 200 word scene that has 90% words of dialogue between two thieves named Molly and Bill who are planning to break into a children's party to steal all the presents. Only include the prose of the scene without any markdown or commentary.
Model response
Messages
Meta
AI
“Bill, this is the lowest point of my criminal career.” “Lower than the inflatable-castle insurance scam?” “That was entrepreneurship. This is larceny with balloons.” “Balloons are festive camouflage, Molly. Nobody suspects a woman in a paper crown.” “I look deranged in a paper crown.” “You look approachable. Predators of joy must look approachable.” “Don’t say ‘predators of joy.’ It sounds like we eat clowns.” “We might, if the buffet’s bad.” “Focus. We get in, lift the presents, get out before the cake knife appears.” “The cake knife?” “Children become feral near buttercream.” “True. What’s our cover?” “You’re Uncle Bill from Swindon.” “I can’t be from Swindon. My knees are too metropolitan.” “Fine. You’re Bill the Magician.” “I know no magic.” “You stole three watches off a vicar during a christening.” “That was sleight of hand.” “Exactly. I’m Molly, frazzled event helper.” “You already look frazzled.” “Thank you, Prince Charming.” “What about the presents table?” “By the window, according to my source.” “Your source being?” “The birthday boy’s older sister. I bribed her with lip gloss and revolutionary rhetoric.” “Promising. Escape route?” “Through the kitchen, over the fence, into the alley.” “With a sack of presents?” “With dignity if possible, dragging if necessary.” “And if a child spots us?” “We improvise.” “Meaning?” “Cry first. Children hate admin.”
created_at
"2026-03-06T12:04:32.398Z"
3/6/2026, 12:04:32 PM
model
"openai/gpt-5.4-20260305"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
504
moderation_latency
(null)
generation_time
11019
tokens_prompt
77
tokens_completion
459
native_tokens_prompt
66
native_tokens_completion
329
native_tokens_completion_images
0
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
"completed"
usage
0.0051
router
(null)
provider_responses
0
endpoint_id
"9ff5625c-403f-4d7f-b895-58ac7295062c"
id
"resp_00ec32e8edbf99410169aac2d0762c819691ca11aa583a5365"
is_byok
false
latency
150
model_permaslug
"openai/gpt-5.4-20260305"
provider_name
"OpenAI"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.0; linux; x64))"
http_referer
(null)
api_type
"completions"
id
"gen-1772798672-qnPtoztJVKHpmcINYSax"
upstream_id
"resp_00ec32e8edbf99410169aac2d0762c819691ca11aa583a5365"
total_cost
0.0051
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenAI"
Evaluation details
Result
Evaluator
Details
Meta Data
51.93%
Matches word count
n/a
neededClean
false
words
216
36.79%
Dialogue to Total Word Ratio
Ratio: 100.00%, Deviation: 10.00%
neededClean
false
wordsTotal
217
wordsDialogue
217
44.3567%