NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-09-01KX41MN5PDVPSX3K7XD896B7Q
Run Details
Model:
GPT-5.6 Luna (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
6.6s
Cost:
$0.005766
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
6.62%
Dialogue to Total Word Ratio
(details)
3.3113%
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
Rain ticked against the van’s windshield while Molly studied the blueprint upside down. “According to this,” Bill whispered, “the birthday party occupies the community hall from noon until three, the presents are stacked beside the magician’s table, and the back door has a faulty latch.” “You’ve been holding that map upside down for ten minutes.” “I’m studying it from the guards’ perspective.” “The guards are six-year-olds.” “Exactly. Unpredictable, heavily caffeinated, and emotionally committed to perimeter defense.” Molly rotated the paper. “Our objective is simple. We enter disguised as entertainers, reach the gift pile, replace every present with something worthless, and leave before the cake is served.” “Something worthless?” “I have prepared eighty-seven wrapped bricks.” Bill glanced into the rear of the van. “Those are not bricks. They’re books.” “Old telephone directories.” “Still educational.” “Children hate education on birthdays.” “They hate vegetables. Education is different.” Molly tapped the blueprint. “The theme is pirates. We dress as pirates, infiltrate the party, and sail away with the treasure.” “We don’t have pirate costumes.” “We have striped shirts, eye patches, and your regrettable beard.” “My beard is magnificent.” “It’s shaped like a frightened squirrel.” “It’s a disguise.” “It’s a cry for help.” Bill pulled a velvet bag from beneath his seat. “I also acquired these.” Molly peered inside. “Plastic swords?” “Authentic pirate cutlasses.” “They bend.” “That’s how you know they’re authentic. Real pirates were safety-conscious.” A gust of rain rattled the van. Molly folded the map and tucked it into her coat. “Phase one,” she said, “we approach the side entrance carrying the cake.” “Why are we carrying the cake?” “Because the cake is our access credential.” “You baked a cake?” “I purchased a cake.” “That explains the box labeled ‘Emergency Birthday Cake.’” “It was the only one left.” “Why does it say ‘Congratulations, Kevin, on Your Retirement’?” “Children don’t read cursive.” “Neither do I.” “Phase two, we charm the host.” “What if the host asks which child we’re related to?” “You’re Uncle Bill.” “I’m nobody’s uncle.” “You are today.” “What’s the child’s name?” “Kevin.” “The retiring one?” “Apparently.” Bill straightened his collar. “And what do I do once we reach the presents?” “Distract the children.” “With what?” “Your ventriloquism.” “I don’t do ventriloquism.” “You do now.” Bill produced a wooden dummy from the glove compartment. Its painted smile was alarmingly wide. “His name is Professor Pickles,” Bill said. “Why does Professor Pickles have a moustache?” “Academic authority.” “He looks like he knows where bodies are buried.” “He’s a respected scholar.” Molly sighed. “You’ll perform until I’ve loaded the gifts.” “And if the children ask for a song?” “I’ll improvise.” “You can’t sing.” “I can mime.” “That’s not singing.” “It is when the audience is under eight.” Bill opened the van door. “What’s the escape route?” “Through the kitchen, over the hedge, across the playground, and into the storm drain.” “We’re stealing presents, not robbing Fort Knox.” “Fort Knox has fewer juice boxes.” They stepped into the rain. Bill tucked Professor Pickles under his arm. “One last question,” he said. “What are we going to do with eighty-seven telephone directories?” Molly smiled. “Give them back as party favors.” “Brilliant.” “Obviously.” “Criminally brilliant.” “Don’t ruin the mood.” From the hall came a shriek of children and the unmistakable sound of a balloon exploding. Bill stopped. “Molly?” “Yes?” “I suddenly feel underqualified.” She handed him the cake. “That’s the first qualification.”
created_at
"2026-07-09T18:18:33.021Z"
7/9/2026, 6:18:33 PM
model
"openai/gpt-5.6-luna-20260709"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1689
moderation_latency
(null)
generation_time
6381
tokens_prompt
77
tokens_completion
1325
native_tokens_prompt
66
native_tokens_completion
950
native_tokens_completion_images
0
native_tokens_reasoning
133
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
0.005766
router
(null)
provider_responses
0
endpoint_id
"eebea444-34a5-4642-a766-cb319471d33a"
id
"resp_0c6c7869e94586fb016a4fe5f918388196a97d4030b6a29de4"
is_byok
false
latency
420
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-1783621113-Nny1v0WynnEnUKII67Ra"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1783621113-cmDSpdDDPdkaYARQbYPG"
upstream_id
"resp_0c6c7869e94586fb016a4fe5f918388196a97d4030b6a29de4"
total_cost
0.005766
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
565
6.62%
Dialogue to Total Word Ratio
Ratio: 77.37%, Deviation: 7.37%
neededClean
false
wordsTotal
570
wordsDialogue
441
3.3113%