NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-03-18-01KM08T9H3BH5EANDWGWB3AKB7
Run Details
Model:
GPT-5.4 Nano
Test:
Dialogue tags
Scenario:
Write 200 words with 90% dialogue
Duration:
3.7s
Cost:
$0.000469
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
0.49%
Matches word count
(details)
99.97%
Dialogue to Total Word Ratio
(details)
50.2332%
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
“Molly, do not step on the squeaky tile again,” Bill whispered. “I swear last time it sounded like a kazoo having a panic attack.” “It wasn’t me, it was the floor’s emotional response to our greatness,” Molly said. “Focus. Children’s party. Presents. Cashmere wrapping paper.” “Gross, not cashmere. Candy canes. Tiny trains. That one glittery spaceship kid promised their cousin they’d never share.” “We share,” Molly said. “Just not with the kid.” “Morality aside, how do we get in without the face-scanning guardian dog?” “We don’t bring the dog,” Molly replied. “We bring the decoy: your mustache.” “My mustache is not a decoy,” Bill hissed. “It’s a lifestyle.” “Lifestyle doesn’t fit the required disguise,” Molly said. “You’re wearing the inflatable clown nose.” “I refuse. People will think I’m funny.” “That’s the goal. We become harmless comedy. Like a balloon animal that steals.” “Balloons don’t steal,” Bill said. “They do if you pop them at the right time.” “Fine. What’s the plan?” “We enter,” Molly said, “we charm the parents with small talk, then we loot the presents during the magic-show distraction.” “And what if the magician sees through our capes?” “Then we accuse him of stealing our dignity,” Molly said. “Now hush. The party’s starting.” “Wait, why are you holding a bag labeled ‘LEGOS’?” “Because it’s polite,” Molly whispered. “Also, it’s heavy.” “That’s not polite. That’s foreboding.”
created_at
"2026-03-18T10:45:43.601Z"
3/18/2026, 10:45:43 AM
model
"openai/gpt-5.4-nano-20260317"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
565
moderation_latency
(null)
generation_time
3603
tokens_prompt
77
tokens_completion
432
native_tokens_prompt
66
native_tokens_completion
365
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
0.00046945
router
(null)
provider_responses
0
endpoint_id
"0c835f2e-c18d-4e8c-b245-e1e3bd08b97f"
id
"resp_01614edb63873a3e0169ba8257b07881948540e7226dc42f4d"
is_byok
false
latency
307
model_permaslug
"openai/gpt-5.4-nano-20260317"
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-1773830743-w4mALaWSrgHswWYVrA4W"
upstream_id
"resp_01614edb63873a3e0169ba8257b07881948540e7226dc42f4d"
total_cost
0.00046945
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenAI"
Evaluation details
Result
Evaluator
Details
Meta Data
0.49%
Matches word count
n/a
neededClean
false
words
227
99.97%
Dialogue to Total Word Ratio
Ratio: 91.27%, Deviation: 1.27%
neededClean
false
wordsTotal
229
wordsDialogue
209
50.2332%