NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-03-09-01KK9PVYVQ29MSS96RRHHQ7CNX
Run Details
Model:
GPT-5.4 (Reasoning)
Test:
Dialogue tags
Scenario:
Write 200 words with 50% dialogue
Duration:
1m 39s
Cost:
$0.085260
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
100.00%
Dialogue to Total Word Ratio
(details)
99.9985%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 200 word scene that has 50% 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 unfolded the crayon-smeared invitation like blueprints for a casino. Balloons tugged at the fence. Beyond the hedge, a clown barked orders at children orbiting a bouncy castle piled high with gifts. Bill crouched beside her, wearing black gloves and a stolen party hat for camouflage. The alley smelled of frosting, wet grass, and criminal ambition. A pony snorted near the gate. Somewhere, a magician misfired confetti and got booed by seven-year-olds. Molly traced the route from gate to present table with one gloved finger, calculating timings, obstacles, and the patrol pattern of a serious grandmother nearby. Molly spoke first. “We go in smiling. Nobody suspects cheerful adults near sheet cake.” “I’ve never looked cheerful. I look like unpaid rent in shoes.” “Perfect. Parents at parties all resemble minor financial disasters.” “What about the clown?” “Bribe him with a balloon dog and professional jealousy.” “And the presents?” “Snatch, sack, vanish through the petting zoo.” “Past the goat? That beast remembers me.” “Good. If it charges, everybody watches the goat.” “That’s either strategy or confession.” “At the piñata signal, we rob innocence before cake arrives.” “Fine, but if a grandmother tackles me, I’m pretending to be the magician’s rabbit again tonight.”
created_at
"2026-03-09T16:28:46.336Z"
3/9/2026, 4:28:46 PM
model
"openai/gpt-5.4-20260305"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
93809
moderation_latency
(null)
generation_time
98281
tokens_prompt
77
tokens_completion
718
native_tokens_prompt
66
native_tokens_completion
5673
native_tokens_completion_images
0
native_tokens_reasoning
5392
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
0.08526
router
(null)
provider_responses
0
endpoint_id
"9ff5625c-403f-4d7f-b895-58ac7295062c"
id
"resp_046eadc7fa67f3710169aef53f24ac8197ab2a1e1e804bc010"
is_byok
false
latency
183
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-1773073726-sg0izjNdHF4Em5BNLGKT"
upstream_id
"resp_046eadc7fa67f3710169aef53f24ac8197ab2a1e1e804bc010"
total_cost
0.08526
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenAI"
Evaluation details
Result
Evaluator
Details
Meta Data
100.00%
Matches word count
n/a
neededClean
false
words
200
100.00%
Dialogue to Total Word Ratio
Ratio: 49.26%, Deviation: 0.74%
neededClean
false
wordsTotal
203
wordsDialogue
100
99.9985%