NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-16-01KHK973JCVCRJ9K6S50468FYN
Run Details
Model:
GPT-5
Test:
Dialogue tags
Scenario:
Write 200 words with 90% dialogue
Duration:
1m 5s
Cost:
$0.043742
Provider:
OpenAI
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
Cake fumes drifted; balloons squeaked; the piƱata judged them silently. "Okay, Bill, target acquired: a glittering fortress of wrapping paper." "Molly, it's a children's party, not fortress; there are cupcakes." "Exactly, cupcakes are guard dogs with frosting; we move smiling." "Our disguises? I still think tiny top hats scream innocence." "And I still think innocence screams when you juggle pineapples." "Fine, no juggling; we mingle, compliment crayons, locate shiny parcels." "About the clown: he's watching me with eyebrows shaped sirens." "Compliment the eyebrows; nobody suspects thieves who admire facial architecture." "Then we execute Plan Sprinkle: distract with confetti, scoop, scoot." "Plan Sprinkle sounds sticky; my pockets remember last Tuesday's sugar." "This time, strategic napkins; we build a portable doily vault." "Is a doily vault bulletproof against icing and moral judgment?" "Absolutely; also, children barter, so trade wisdom for ribbon intelligence." "What wisdom? Don't microwave fish? Always label leftovers? Hug responsibly?" "No, profound things: gravity exists, and grownups secretly fear glitter." "Glitter is forever; it's the diamond heist of craft supplies." "Focus; we steal all the presents, like philanthropists in reverse." "If he vanishes us, I'm haunting the cake with poltergeisting." "Deal; on three: one, two, compliment, sprinkle, scoop, scoot, retreat."
created_at
"2026-02-16T13:12:17.888873+00:00"
2/16/2026, 1:12:17 PM
model
"openai/gpt-5-2025-08-07"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
34529
moderation_latency
(null)
generation_time
64875
tokens_prompt
77
tokens_completion
1039
native_tokens_prompt
66
native_tokens_completion
4366
native_tokens_completion_images
0
native_tokens_reasoning
4032
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
0.0437425
router
(null)
provider_responses
0
id
"resp_071aa9dfcf7ed3120169931770496c8194af90fabc6d1ea2f5"
status
200
is_byok
false
latency
163
endpoint_id
"7c2f859a-7890-4e8e-b1de-1cd1c0a800b4"
provider_name
"OpenAI"
model_permaslug
"openai/gpt-5-2025-08-07"
api_type
"completions"
id
"gen-1771247472-kTFHY8TTdEtnILgNU3g0"
upstream_id
"resp_071aa9dfcf7ed3120169931770496c8194af90fabc6d1ea2f5"
total_cost
0.0437425
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenAI"
Evaluation details
Result
Evaluator
Details
Meta Data
100.0000%
Matches word count
n/a
neededClean
false
words
200
93.9413%
Dialogue to Total Word Ratio
Ratio: 95.00%, Deviation: 5.00%
neededClean
false
wordsTotal
200
wordsDialogue
190
96.9707%