NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-03-18-01KM083V0F16DYZ8CRNHXVA48P
Run Details
Model:
GPT-5.4 Nano
Test:
Dialogue tags
Scenario:
Write 200 words with 10% dialogue
Duration:
2.1s
Cost:
$0.000477
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
0.0000%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 200 word scene that has 10% 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 crouched beside the toy-store van, adjusting the suction cups on her gloves with the seriousness of a surgeon. Bill hovered over the shopping cart they’d “borrowed,” which rattled with dignity it didn’t earn. Outside, a bouncy castle quivered like a nervous jellyfish under a balloon arch that screamed PARTY in letters shaped like cartoon hearts. “Okay,” Molly whispered, tracing a map of the yard in the dirt. “We slip through the gate, neutralize the cookie jar alarm, and—” Bill’s grin widened. “And steal every last present. Like a seasonal menace.” “Less menace,” Molly said, pausing to listen. From the house came a muffled countdown and someone singing the birthday song at the wrong speed. “The kids are practicing,” Bill added, like that was a real job requirement. Molly snapped her fingers and held up a spool of rainbow ribbon. “We’ll distract them with confetti camouflage.” Bill nodded, then pointed at the security camera. “What about that?” Molly tapped her earpiece, then revealed a tiny plastic duck with googly eyes. “Duck-based disruption device.” Bill snorted. “That’s not a real thing.” Molly’s eyes glittered. “It’s absolutely a real thing. Ask the duck.” “Great,” Bill said. “I’m stealing presents from a party full of toddlers.” Molly leaned closer, voice steady and conspiratorial. “Exactly. We’re professionals.” Bill adjusted his mask, which was already smudged with frosting he hadn’t stolen yet. “Let’s do this before the magician starts.” Molly raised the duck. “On three.” “One,” Bill said. “Two,” Molly corrected. “Three,” they whispered together, and the yard lights flickered like the house was laughing.
created_at
"2026-03-18T10:33:27.838Z"
3/18/2026, 10:33:27 AM
model
"openai/gpt-5.4-nano-20260317"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
464
moderation_latency
(null)
generation_time
1903
tokens_prompt
77
tokens_completion
449
native_tokens_prompt
66
native_tokens_completion
371
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
0.00047695
router
(null)
provider_responses
0
endpoint_id
"0c835f2e-c18d-4e8c-b245-e1e3bd08b97f"
id
"resp_0ada9762779737760169ba7f77e6b8819583387a23716dc39e"
is_byok
false
latency
209
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-1773830007-7mvsAio2lGAP2WK1o0P2"
upstream_id
"resp_0ada9762779737760169ba7f77e6b8819583387a23716dc39e"
total_cost
0.00047695
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
260
0.00%
Dialogue to Total Word Ratio
Ratio: 29.89%, Deviation: 19.89%
neededClean
false
wordsTotal
261
wordsDialogue
78
0.0000%