NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-03-18-01KM0ACABKMQBVY5333RP0GWV2
Run Details
Model:
GPT-5.4 Nano
Test:
Dialogue tags
Scenario:
Write 200 words with 10% dialogue
Duration:
1.9s
Cost:
$0.000424
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
8.43%
Dialogue to Total Word Ratio
(details)
4.2141%
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 adjusted her striped beanie while the moonlight slid across the suburban lawn like it was trying to mind its own business. Bill held a clipboard that definitely didn’t belong to any adult with a job, because it was filled with stick-figure diagrams of mousetrap placements and one horrifying drawing labeled “CONFIDENCE.” From the living room came muffled squeals, the kind generated when someone discovers confetti or a new dinosaur-themed sticker sheet. Molly glanced at the front door, then at Bill. “We’re not stealing joy. We’re relocating it.” Bill nodded solemnly, then whispered, “Relocating joy to our secret lair museum.” Molly smirked and slid a thin tool into the lock with the grace of a person who’d watched three YouTube tutorials and ignored all warnings. The latch clicked. Somewhere inside, a balloon bobbed like it knew too much. Bill inhaled, then tried to look casual, which only made him look like a baked potato in a tuxedo. “Remember: no touching the cupcake bar,” Molly said, stepping through first. “Unless the cupcakes have valuables,” Bill replied, already reaching for a clipboard-shaped flourish. Molly glanced back, eyes narrowing. “You can’t steal sprinkles.” “Bold of you to assume the sprinkles aren’t diamonds,” Bill whispered. They slipped deeper into the hallway, listening to the children laugh like tiny alarms, and Molly quietly reminded him, “If anyone asks, we’re… special delivery.” Bill saluted with two fingers. “Special delivery of stolen presents. Now lead the way, maestro.”
created_at
"2026-03-18T11:13:02.849Z"
3/18/2026, 11:13:02 AM
model
"openai/gpt-5.4-nano-20260317"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
426
moderation_latency
(null)
generation_time
1504
tokens_prompt
77
tokens_completion
402
native_tokens_prompt
66
native_tokens_completion
329
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
0.00042445
router
(null)
provider_responses
0
endpoint_id
"0c835f2e-c18d-4e8c-b245-e1e3bd08b97f"
id
"resp_01d166c411f598d50169ba88bee5a881949948d1c982da1543"
is_byok
false
latency
157
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-1773832382-6HhDax97CUHskQcEeDat"
upstream_id
"resp_01d166c411f598d50169ba88bee5a881949948d1c982da1543"
total_cost
0.00042445
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
241
8.43%
Dialogue to Total Word Ratio
Ratio: 22.54%, Deviation: 12.54%
neededClean
false
wordsTotal
244
wordsDialogue
55
4.2141%