NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-03-06-01KK1GHTS9523YE42BZVJJ0C9E
Run Details
Model:
GPT-5.4
Test:
Dialogue tags
Scenario:
Write 200 words with 90% dialogue
Duration:
8.0s
Cost:
$0.004725
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
75.16%
Matches word count
(details)
36.79%
Dialogue to Total Word Ratio
(details)
55.9718%
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, tell me again why we’re robbing a children’s party.” “Because, Molly, rich parents buy expensive toys, and expensive toys are merely jewels with batteries.” “That is the worst philosophy I’ve ever heard, and I once dated a man who believed soup was a beverage.” “Focus. Balloons at noon, cake at one, screaming by one-ten. We go in during the magician.” “The magician?” “No one watches the gift table when a man pulls a pigeon out of his sleeve.” “What if the pigeon watches us?” “Then we bribe it with a sausage roll.” “Bill, children are sharp. Tiny, sticky detectives.” “Exactly why we wear disguises.” “I’m not dressing as a clown. If they catch me in face paint, I’ll throw myself to the authorities.” “Not clowns. Parents. The perfect camouflage: exhausted, underdressed, carrying coffee.” “I can do exhausted. I was born exhausted.” “You’ll drift by the presents, I’ll create a diversion near the bouncy castle.” “How big a diversion?” “Large enough that every mother says, ‘Oh my God, Bill.’” “That usually means fire.” “Only a little.” “And the escape?” “Through the garden gate with the loot in a stroller.” “A stroller?” “No one questions a stroller.” “What if there’s an actual baby in it?” “Then we abort.” “The job?” “No, Molly. The baby.”
created_at
"2026-03-06T12:04:27.571Z"
3/6/2026, 12:04:27 PM
model
"openai/gpt-5.4-20260305"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
482
moderation_latency
(null)
generation_time
7911
tokens_prompt
77
tokens_completion
406
native_tokens_prompt
66
native_tokens_completion
304
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
0.004725
router
(null)
provider_responses
0
endpoint_id
"9ff5625c-403f-4d7f-b895-58ac7295062c"
id
"resp_04e2722521df21ef0169aac2cba3288193a09f4079fe30c4ee"
is_byok
false
latency
165
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-1772798667-TsalKEa3Y1g2O3wdQk0Z"
upstream_id
"resp_04e2722521df21ef0169aac2cba3288193a09f4079fe30c4ee"
total_cost
0.004725
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenAI"
Evaluation details
Result
Evaluator
Details
Meta Data
75.16%
Matches word count
n/a
neededClean
false
words
213
36.79%
Dialogue to Total Word Ratio
Ratio: 100.00%, Deviation: 10.00%
neededClean
false
wordsTotal
214
wordsDialogue
214
55.9718%