NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-09-01KX411QY3PZHHG2QV93ZKZRRP
Run Details
Model:
GPT-5.6 Luna
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
5.4s
Cost:
$0.004236
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
43.38%
Matches word count
(details)
79.95%
Dialogue to Total Word Ratio
(details)
61.6652%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 30% 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 blueprint across the laundromat table, pinning its corners with four suspiciously heavy socks. “According to this,” she whispered, “the birthday party begins at noon, the cake arrives at twelve-thirty, and the presents are stacked beneath the balloon arch.” Bill peered at the drawing. “Why is the balloon arch guarded by a cartoon dinosaur?” “That’s not a guard. That’s the entertainment.” “It has teeth.” “It’s wearing a party hat.” “So did my uncle at Christmas, and he still bit three people.” Outside, a delivery van rattled past, scattering pigeons and one elderly man who had apparently been waiting for a bus since 1987. Molly tapped the blueprint with a gloved finger. “The children’s entrance is on the east side. We slip in disguised as entertainers, locate the gift table, fill these bags, and leave before the magician starts.” “What are we disguised as?” “Clowns.” Bill stared at her. Molly stared back. “Fine,” he said. “But if anyone asks, I’m a balloon sculptor.” “You can’t sculpt balloons.” “I can make a tube.” “That’s not sculpture. That’s evidence.” They crossed the street at precisely eleven fifty-seven, carrying two enormous sacks and wearing matching red wigs. Molly had painted a smile on her face. Bill had painted one too, but his smile wandered toward his ear, giving him the expression of a man who had just remembered a very small crime. At the side gate, a woman with a clipboard blocked their path. “Names?” she asked. “Molly and Bill,” said Molly. “Performers?” “Absolutely,” said Bill. “We specialize in festive shapes.” The woman glanced at his empty hands. “Where are your balloons?” Bill patted his coat. A balloon squeaked somewhere inside it. “Sensitive equipment,” he said. She waved them through. The backyard seethed with children, frosting, and noise. A boy in a pirate hat chased a girl dressed as a fairy around a plastic slide. Near the table, a toddler sat solemnly beside the presents, guarding them with the grim authority of a tiny customs officer. Molly nudged Bill. “Stage one.” Bill approached the balloon arch and pulled out a red balloon. He inflated it with impressive confidence, twisted it twice, and produced something resembling a startled intestine. “What is it?” Molly murmured. “A dachshund.” “It has no head.” “It’s abstract.” Before she could answer, the birthday child spotted them. “Make me a sword!” “Of course,” Bill said. Within minutes, he was surrounded by eager children demanding swords, hats, giraffes, and one remarkably detailed representation of a grandmother. Molly drifted toward the gift table, but every time she reached for a package, another child handed Bill a balloon animal and asked for something more difficult. “Bill,” she hissed, “we’re here to steal presents.” “I know,” he whispered, wrestling a balloon into the shape of a swan. “But apparently I’m good at this.” The magician arrived early, saw Bill’s creations, and hired him on the spot. By sunset, Molly and Bill left with no presents, three unpaid balloon orders, and a thank-you card from the birthday child. Inside it was a crayon drawing of Bill’s abstract dachshund. He framed it.
created_at
"2026-07-09T18:08:13.26Z"
7/9/2026, 6:08:13 PM
model
"openai/gpt-5.6-luna-20260709"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
671
moderation_latency
(null)
generation_time
5045
tokens_prompt
77
tokens_completion
924
native_tokens_prompt
66
native_tokens_completion
695
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)
num_fetches
(null)
web_search_engine
(null)
origin
"https://ncbench.com/"
is_byok
false
finish_reason
"stop"
native_finish_reason
"completed"
service_tier
"default"
usage
0.004236
router
(null)
provider_responses
0
endpoint_id
"eebea444-34a5-4642-a766-cb319471d33a"
id
"resp_0239c81543b99e44016a4fe38d55a4819798587cf832e77b66"
is_byok
false
latency
389
model_permaslug
"openai/gpt-5.6-luna-20260709"
provider_name
"OpenAI"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.18.0; linux; x64))"
http_referer
(null)
request_id
"req-1783620493-CvESPVI9N93ESEkZm7wM"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1783620493-1uN3baBxeYrkiMIGiEgf"
upstream_id
"resp_0239c81543b99e44016a4fe38d55a4819798587cf832e77b66"
total_cost
0.004236
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenAI"
response_cache_source_id
(null)
data_region
"global"
Evaluation details
Result
Evaluator
Details
Meta Data
43.38%
Matches word count
n/a
neededClean
false
words
517
79.95%
Dialogue to Total Word Ratio
Ratio: 32.12%, Deviation: 2.12%
neededClean
false
wordsTotal
520
wordsDialogue
167
61.6652%