NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-23-01M36RF5TNR60RFXKZGA6FAGB2
Run Details
Model:
GPT-6 Luna (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
21.1s
Cost:
$0.000755
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
0.0000%
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
“Are you sure this is the party?” Molly asked, peering through the van’s rain-streaked window. Bill checked the invitation. “It says, ‘Toby’s sixth birthday, Saturday, two o’clock.’ It is Saturday. It is two o’clock. And there’s a banner that says TOBY.” “It also says ‘No grown-ups without a grown-up.’” “That’s a rule for guests.” “We’re not guests.” “We’re professionals.” “Professionals don’t wear matching fake moustaches.” “They were on sale in a two-pack.” Molly lowered her binoculars. “The presents are stacked beside the cake.” “Excellent.” “Under a sign that says PLEASE LEAVE PRESENTS HERE.” “Even better. They’ve organized the loot for us.” “They’ve organized birthday presents for a six-year-old.” “Which is why we’ll be quick, discreet, and completely invisible.” Bill climbed out of the van wearing a purple party hat. Molly stared at it. “What?” he said. “It’s camouflage.” “It has a spinning propeller.” “Children love propellers.” “So do toddlers. That’s not the same as blending in.” Bill tugged the hat off. “Fine. We enter as entertainers.” “What do you entertain?” “I can juggle.” “With what?” “Three oranges.” “Those are our lunch.” “I can juggle two.” Molly considered the house. From inside came shrieks, laughter, and the unmistakable sound of a child discovering a kazoo. “We walk in, take every present, and walk out. No improvising.” “Absolutely. I’ll be the distraction.” “You are always the distraction.” “I’m a natural.” “You once distracted a bank teller by asking if the vault was heated.” “And she answered! Very friendly woman.” The front door opened. A little boy shot out, clutching a cardboard crown. He stopped beside the van and peered at them. “Are you the magicians?” he asked. Bill smiled. “We can be.” “Can you make my uncle disappear? He keeps doing impressions.” Molly leaned toward Bill. “That sounds like a paying job.” The boy pointed at the presents. “Those are for me. But you can have a cupcake if you come inside.” Bill glanced at the stack of gifts. Molly glanced at Bill. For a moment, the whole elaborate plan hung between them, flimsy as Bill’s moustache. “Cupcake first,” Molly said. “Then we steal the presents?” Bill whispered. “Then we assess the cupcake.” They followed the boy inside. A woman in a paper crown handed them each a cupcake and asked if they were the entertainers. “Professional magicians,” Bill said. “Formerly,” Molly added. A child shoved a kazoo into Bill’s hand. “Make it vanish!” Bill looked at Molly. She took the kazoo, tucked it into her sleeve, and produced it from behind the child’s ear. The room erupted. “Can you do the presents next?” the boy asked. Molly looked at the bright pile, at the birthday boy’s delighted face, and at Bill already reaching for a second cupcake. “Certainly,” she said. “We’ll make them all disappear.” The children gasped. “Into your bedroom,” Molly added. “Before anyone trips over them.” Bill nodded solemnly. “A very advanced illusion.” They spent the next hour carrying gifts to the bedroom, one at a time. When they finally left, Molly had frosting on her coat, Bill had a paper crown on his head, and neither had stolen a thing. “Successful operation?” Bill asked. Molly considered the empty cupcake tray. “We got away with dessert.”
created_at
"2026-09-23T09:10:10.53Z"
9/23/2026, 9:10:10 AM
model
"openai/gpt-6-luna-20260922"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
6264
moderation_latency
(null)
generation_time
21087
tokens_prompt
75
tokens_completion
1079
native_tokens_prompt
66
native_tokens_completion
1497
native_tokens_completion_images
0
native_tokens_reasoning
732
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
0.0007551
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1790154610-6WpSfAssjOTPOaxT8OJe"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790154610-cHZKtbYECESmgvfLqUUW"
upstream_id
"resp_0c23e489af8b3343016ab39772a20087d1a066be206135c8bf"
provider_responses
0
endpoint_id
"05e94e02-b9c2-4bbb-ba55-4082ee9ad687"
id
"resp_0c23e489af8b3343016ab39772a20087d1a066be206135c8bf"
is_byok
false
latency
709
model_permaslug
"openai/gpt-6-luna-20260922"
provider_name
"OpenAI"
status
200
total_cost
0.0007551
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenAI"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
537
0.00%
Dialogue to Total Word Ratio
Ratio: 51.75%, Deviation: 18.25%
neededClean
false
wordsTotal
543
wordsDialogue
281
0.0000%