NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-09-01KX409Q8XMF8H2Q9MAQP10VZZ
Run Details
Model:
GPT-5.6 Luna
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
5.2s
Cost:
$0.004512
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
97.63%
Matches word count
(details)
0.02%
Dialogue to Total Word Ratio
(details)
48.8241%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 70% 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
“Tell me again why we’re robbing a children’s party?” “Because, Bill, the presents are small, portable, and wrapped by people who still believe tape is a security system.” “That’s not a plan. That’s a complaint.” “It’s the beginning of a plan. We enter during the magician’s act, distract the adults, and remove the gifts.” “Remove?” “Steal.” “You said remove.” “I’m trying to make us sound professional.” “We’re breaking into a room full of seven-year-olds.” “Exactly. They’ll assume we’re entertainers.” “You’re wearing a black turtleneck and one sock.” “The sock is tactical.” “It has a duck on it.” “Ducks are stealthy. They float silently.” “They quack.” “Only when provoked.” Molly crouched behind the hedge outside the community hall and peered through the window. Inside, balloons bobbed beneath a banner reading HAPPY BIRTHDAY, JASPER! A mountain of presents occupied the corner, guarded by a solemn child in a paper crown. “Look,” she whispered. “The crown guard.” “He’s eight.” “His posture says trained.” “His posture says he swallowed a marble.” “Focus. We need disguises.” Bill opened his suitcase. “I brought two options: clowns or catering staff.” “Those are both terrible.” “The clown costume has pockets.” “The catering uniform has pockets.” “Then we’re agreed.” “Which one?” “Both.” Five minutes later, Molly wore a red clown wig over a catering cap, while Bill had tied an apron around a purple clown suit. “You look like a kitchen accident,” Molly said. “You look like a clown who lost a bet.” “That’s our cover. We’re the entertainment.” “We are not entertaining.” “We don’t need to be good.” They slipped through the side door and entered just as a magician pulled three rubber chickens from a hat. “Ladies and gentlemen,” Molly announced, clapping loudly, “the second act has arrived!” A dozen children turned. “Are you clowns?” asked a girl with frosting on her nose. “Only professionally,” said Bill. “Can you make my brother disappear?” “Temporarily,” Molly said. “Possibly into a cupboard.” “Don’t promise that,” Bill muttered. The magician glared at them. “I’m performing.” “Wonderful,” said Molly. “Keep doing that.” Bill drifted toward the presents, humming. The crown guard stood in his path. “Password,” said the boy. Bill froze. “Birthday?” “Wrong.” “Cake?” “Wrong.” “Your mother’s maiden name?” The boy narrowed his eyes. “Nice try.” Molly swooped in, holding out a balloon animal. “Is this your mother?” The boy stared at the crooked balloon dog. “No.” “Then guard it.” He accepted the animal, distracted. Bill began sliding presents into a catering cart. “Why is that box moving?” asked a parent. “Interactive gift,” Bill said. “It’s making a rattling noise.” “Very interactive.” Molly grabbed the cart. “We’re leaving.” “Already?” “The children are becoming suspicious.” “They’re children,” Bill said. “They were suspicious before we arrived.” They burst through the door, pushed the cart three feet, and stopped. A tiny girl stood blocking the path, clutching a teddy bear. “You forgot my present,” she said. Molly glanced at the cart, then at the girl. “Which one is yours?” “All of them.” Bill sighed. “I knew we should’ve robbed a dentist.”
created_at
"2026-07-09T17:55:06.149Z"
7/9/2026, 5:55:06 PM
model
"openai/gpt-5.6-luna-20260709"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
498
moderation_latency
(null)
generation_time
4853
tokens_prompt
77
tokens_completion
1039
native_tokens_prompt
66
native_tokens_completion
741
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
0.004512
router
(null)
provider_responses
0
endpoint_id
"eebea444-34a5-4642-a766-cb319471d33a"
id
"resp_03144fb73df40ae4016a4fe07a3ab48193b2edce0440621ffe"
is_byok
false
latency
257
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-1783619706-1nfCfkK05ArBKLULTwxF"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1783619706-z45I048qaZy5lTN6mJJB"
upstream_id
"resp_03144fb73df40ae4016a4fe07a3ab48193b2edce0440621ffe"
total_cost
0.004512
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
97.63%
Matches word count
n/a
neededClean
false
words
507
0.02%
Dialogue to Total Word Ratio
Ratio: 56.97%, Deviation: 13.03%
neededClean
false
wordsTotal
509
wordsDialogue
290
48.8241%