NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-03-18-01KM08FY1VMQQWNSCN1NMXGQJX
Run Details
Model:
GPT-5.4 Nano
Test:
Dialogue tags
Scenario:
Write 200 words with 90% dialogue
Duration:
27.7s
Cost:
$0.000471
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.11%
Dialogue to Total Word Ratio
(details)
0.0531%
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, are we seriously doing a children’s party? With balloons?” Bill whispered, gripping the sack like it owed him money. “Shh. We’re professionals. Besides, stealing presents is basically charity,” Molly said. “Charity?” Bill blinked. “For us. The kids won’t even remember what they got. We’ll remember forever. In a trunk. Under my bed.” “Your bed smells like dust and poor decisions,” Molly muttered. “My bed is an artisanal ecosystem.” Bill leaned toward the door. “Okay, step one: how do we get in?” “Easy,” Molly said, flashing a tiny lockpick set. “We say we’re the surprise balloon delivery.” Bill stared. “You can’t just say that.” “Watch me,” Molly said. “Hey, kids, I’m—” “We’re not talking to kids,” Bill cut in. “Right, right. We’re talking to the adult who’s pretending not to notice the sliding shadow.” Bill swallowed. “What adult?” “The one with the ‘I can totally handle sugar’ face,” Molly replied. “The cookie table supervisor. I’ll distract them with enthusiasm.” Bill whispered, “You? Enthusiasm?” “Yes. I can perform joy at gunpoint.” “There’s no gun.” “Exactly. It’s more impressive.” Bill adjusted his gloves. “Fine. And step two?” “We grab the presents, not the cake,” Molly said. “Cake is a trap. Frosting is evidence.” Bill nodded. “And step three?” Molly smiled. “We leave before the piñata starts swinging like a lawsuit.” Bill exhaled. “This is insane.” “It’s hilarious,” Molly corrected, and they slipped toward the door.
created_at
"2026-03-18T10:40:04.166Z"
3/18/2026, 10:40:04 AM
model
"openai/gpt-5.4-nano-20260317"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3037
moderation_latency
(null)
generation_time
27684
tokens_prompt
77
tokens_completion
420
native_tokens_prompt
66
native_tokens_completion
366
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
0.0004707
router
(null)
provider_responses
0
endpoint_id
"0c835f2e-c18d-4e8c-b245-e1e3bd08b97f"
id
"resp_025a7c91687042a70169ba8104d6b88193964778314cffdb4f"
is_byok
false
latency
2079
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-1773830404-6NR3HIVxv6gw9LqXzOu5"
upstream_id
"resp_025a7c91687042a70169ba8104d6b88193964778314cffdb4f"
total_cost
0.0004707
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenAI"
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
233
0.11%
Dialogue to Total Word Ratio
Ratio: 73.82%, Deviation: 16.18%
neededClean
false
wordsTotal
233
wordsDialogue
172
0.0531%