NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-09-01KX41CV56RE0AA32JD5A9MPK1
Run Details
Model:
GPT-5.6 Luna (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
6.3s
Cost:
$0.005196
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
0.0012%
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
Molly crouched behind the hedge and peered through the kitchen window. Inside, balloons bobbed beneath the ceiling, a tower of cupcakes leaned dangerously beside the sink, and a dozen children shrieked around a magician wearing a purple cape. “Report,” Bill whispered. “The target is heavily guarded.” “By whom?” “Children.” Bill frowned. “Armed?” “One has a plastic sword. Another is holding a juice box like he knows how to use it.” “Then we proceed with caution.” “You said that last time before falling through a conservatory roof.” “That roof was structurally overconfident.” Molly checked the map spread across her knees. It was actually a coloring page featuring a dinosaur, but she had drawn arrows over the dinosaur’s stomach. “The presents are stacked beneath the staircase. We enter through the side door, cross the laundry room, bypass the cake table, and extract the packages through the garden gate.” “Why not use the front door?” “Because the front door is where normal people go.” “And we’re thieves.” “Exactly. We have standards.” Bill adjusted his black turtleneck. “What’s the distraction?” “I brought a remote-controlled mouse.” “Excellent.” “It’s pink, wears a tutu, and plays ‘The Wheels on the Bus.’” “Less excellent.” “It was the only one at the novelty shop.” “Fine. We release the mouse. The children chase it. We take the presents.” “What if they catch it?” “Then we improvise.” “What if the birthday child cries?” Molly paused. “We’re stealing from a child?” “We’re stealing presents from several children. That’s different. It’s practically organized.” “That does not make it better.” “It makes it professional.” A cheer erupted inside. The magician had produced a rabbit from his hat, although the rabbit appeared to be a stuffed toy with a price tag still attached. “Now,” Molly whispered. “Go.” Bill nudged open the side door. It squeaked. Every child in the house went silent. The magician slowly turned. “Would anyone like to see another trick?” A small boy pointed toward the doorway. “There are burglars!” Bill stepped inside and raised both hands. “Technically, we’re party consultants.” Molly entered behind him. “Specializing in gift redistribution.” The birthday girl narrowed her eyes. “You’re here to steal my presents.” Bill glanced at Molly. “Children are alarmingly perceptive.” “Give us one good reason we shouldn’t call the grown-ups,” said the girl. Molly looked at the crooked cake, the exhausted magician, and the mountain of brightly wrapped boxes. “Because we can improve this operation.” “How?” “We’ll steal the presents, hide them, and make the magician find them.” The magician brightened. “I’ve been looking for a new finale.” Bill leaned toward Molly. “That is not stealing.” “It is if we keep one.” The girl considered this. “You can keep the empty box.” Bill sighed. “Fine. But we’re taking the biggest one.” “It’s full of socks,” she said. Molly smiled. “A flawless victory.” Five minutes later, the thieves vanished through the garden gate carrying an empty cardboard box, while the children applauded their escape. Inside, the magician pulled the real presents from beneath the sofa. Bill trudged beside Molly. “We’ve become entertainers.” “No,” Molly said. “We’ve become criminals with boundaries.” “Do boundaries pay?” Molly opened the box. A pair of enormous wool socks tumbled out. “Apparently,” she said, “in installments.”
created_at
"2026-07-09T18:14:17.01Z"
7/9/2026, 6:14:17 PM
model
"openai/gpt-5.6-luna-20260709"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1476
moderation_latency
(null)
generation_time
5698
tokens_prompt
77
tokens_completion
1038
native_tokens_prompt
66
native_tokens_completion
855
native_tokens_completion_images
0
native_tokens_reasoning
111
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
0.005196
router
(null)
provider_responses
0
endpoint_id
"eebea444-34a5-4642-a766-cb319471d33a"
id
"resp_0ff09d32660a49ae016a4fe4f91b648195b575a13825cb005c"
is_byok
false
latency
236
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-1783620857-H5rxwaUQ7h1sryi8wYyA"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1783620857-brTQUjuh9WUiVBFMJ6kG"
upstream_id
"resp_0ff09d32660a49ae016a4fe4f91b648195b575a13825cb005c"
total_cost
0.005196
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
0.00%
Matches word count
n/a
neededClean
false
words
536
0.00%
Dialogue to Total Word Ratio
Ratio: 55.39%, Deviation: 14.61%
neededClean
false
wordsTotal
538
wordsDialogue
298
0.0012%