NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-03-06-01KK1HNZMK8HRNQKKWJBBPRAA6
Run Details
Model:
GPT-5.4 (Reasoning)
Test:
Dialogue tags
Scenario:
Write 200 words with 90% dialogue
Duration:
9.3s
Cost:
$0.004770
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
60.28%
Matches word count
(details)
94.26%
Dialogue to Total Word Ratio
(details)
77.2675%
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, where else do you find thirty gifts in one room and absolutely no security besides a clown with asthma?” “I hate clowns. They’re like if lies learned balloon animals.” “Focus. We go in disguised as entertainers.” “I can juggle suspicion and resentment. Does that count?” “You’ll be Princess Sparkle. I’ll be Captain Bouncy.” “That sounds less like a disguise and more like a lawsuit.” “Parents see foam hats and confidence, they stop asking questions.” “What about the presents?” “We sweep them into the cake trolley.” “The cake trolley?” “No one searches cake. It’s a sacred object.” “Bill, children search cake with their hands.” “Fine. Backup plan: giant gift sack.” “That’s Santa. In June. Bold, but seasonally confusing.” “Confusion is our ally.” “Our ally usually gets us chased by dogs.” “Not this time. I bribed the dog walker.” “There’s a dog walker at a party?” “There is when I invent one to sound prepared.” “Marvelous. So we’re underdressed, overconfident, and one balloon pop from prison.” “Exactly. A classic Molly-and-Bill operation.” “If a child asks me to make a rabbit from a napkin, I’m surrendering.” “If a child bites me, I’m keeping the bicycle.” They grinned like saints planning arson and reached for the glitter.
created_at
"2026-03-06T12:24:12.186Z"
3/6/2026, 12:24:12 PM
model
"openai/gpt-5.4-20260305"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
543
moderation_latency
(null)
generation_time
9218
tokens_prompt
77
tokens_completion
415
native_tokens_prompt
66
native_tokens_completion
307
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
0.00477
router
(null)
provider_responses
0
endpoint_id
"9ff5625c-403f-4d7f-b895-58ac7295062c"
id
"resp_0453ad10347a49890169aac76c3fd8819596299de9e0490cd1"
is_byok
false
latency
286
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-1772799852-6KDDJnhy8SadcRTZC5Rs"
upstream_id
"resp_0453ad10347a49890169aac76c3fd8819596299de9e0490cd1"
total_cost
0.00477
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenAI"
Evaluation details
Result
Evaluator
Details
Meta Data
60.28%
Matches word count
n/a
neededClean
false
words
215
94.26%
Dialogue to Total Word Ratio
Ratio: 94.93%, Deviation: 4.93%
neededClean
false
wordsTotal
217
wordsDialogue
206
77.2675%