NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-16-01KHKB46VY2YPE9Y82KTV2FFX9
Run Details
Model:
GPT-5
Test:
Dialogue tags
Scenario:
Write 200 words with 10% dialogue
Duration:
52.7s
Cost:
$0.032302
Provider:
OpenAI
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 flattened the blueprint over a frosting-stained tablecloth stolen from last week’s clown college yard sale. She had annotated the bouncy castle like it was a bank vault, tracing airflow arrows and the likely tantrum epicenters. Bill weighed a corner with a rubber duck grenade that only squeaked menacingly. "We’re robbing a birthday, Bill." Molly said it like reciting a saint’s name before a sacred folly. Bill adjusted his tie, a streamered noose of confidence, and checked the watch he couldn’t stop resetting. "Relax; kids fear spreadsheets, Molly." He flourished a color-coded schedule labeled Diversion by Cupcake. The plan hinged on a magician’s rabbit, two gluten-free decoys, and a fog machine borrowed from a haunted hayride. Molly rehearsed smiles calibrated to aunt-resistant sincerity while Bill counted steps between gift pile and exit like a cat burglar with asthma. "Presents first, cake never, always." She tapped a glitter star: the getaway path through a maze of juice boxes. The door rattled with distant kazoo thunder. Confetti drifted under it like heraldic dandruff. Bill pocketed a whoopee cushion detonator. "Wear socks; bounce houses judge." He wiped frosting from the blueprint, and their fingerprints turned into sprinkles. The piñata watched, blindfolded, already betrayed.
created_at
"2026-02-16T13:45:27.467316+00:00"
2/16/2026, 1:45:27 PM
model
"openai/gpt-5-2025-08-07"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3385
moderation_latency
(null)
generation_time
52297
tokens_prompt
77
tokens_completion
1137
native_tokens_prompt
66
native_tokens_completion
3222
native_tokens_completion_images
0
native_tokens_reasoning
2880
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
0.0323025
router
(null)
provider_responses
0
id
"resp_07b4a5026864650c0169931f428584819592817d16189efd84"
status
200
is_byok
false
latency
148
endpoint_id
"7c2f859a-7890-4e8e-b1de-1cd1c0a800b4"
provider_name
"OpenAI"
model_permaslug
"openai/gpt-5-2025-08-07"
api_type
"completions"
id
"gen-1771249474-kbASihJJrmT1zIcmiuei"
upstream_id
"resp_07b4a5026864650c0169931f428584819592817d16189efd84"
total_cost
0.0323025
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenAI"
Evaluation details
Result
Evaluator
Details
Meta Data
99.7443%
Matches word count
n/a
neededClean
false
words
204
100.0000%
Dialogue to Total Word Ratio
Ratio: 9.80%, Deviation: 0.20%
neededClean
false
wordsTotal
204
wordsDialogue
20
99.8722%