NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-23-01M36P0ZMMDKRDX9F9NDS1VXZC
Run Details
Model:
GPT-6 Luna (Reasoning, Medium)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
33.2s
Cost:
$0.001849
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
100.00%
Dialogue to Total Word Ratio
(details)
100.0000%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 50% 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 spread the party invitation across a laundromat table, weighing its corners with two stale muffins and one exceptionally nervous teaspoon from the vending machine. “First, we enter disguised as responsible adults, then locate the presents, then leave before anyone asks why our gift is a sack of suspicious potatoes.” Bill examined the invitation’s glittery unicorn map, turning it sideways until the birthday cake looked like a compass and the clown resembled a legal threat. “Responsible adults bring presents, Molly. We should carry one enormous box, preferably labeled educational, so nobody suspects the contents are already gone from the building.” Molly drew a diagram on a napkin, featuring the gate, the garden shed, and a dragon marked definitely not a volunteer, probably someone’s uncle. “Entry is simple: I will charm the parents with small talk, while you distract the children using your world famous balloon animals and mild panic.” Bill polished his one balloon, which had survived three birthdays, a bicycle accident, and an unfortunate encounter with a particularly ambitious hedgehog named Sergeant Prickles. “Last time I made a balloon dog, it looked like a tax auditor, and the children demanded receipts before giving me applause or their trust.” Outside, the party house glittered with paper stars; through its windows, a dozen children chased a plastic crown around the living room like tiny lawyers. “Forget balloons. We enter through the side gate, blend with the grownups, and use the phrase, We are here for the presents, and unpaid overtime.” Bill frowned at the word presents, because it sounded less like a plan and more like a confession delivered by a nervous parrot in court. “Could we steal only the boring gifts? Socks, puzzle books, perhaps a sweater knitted by someone who considers beige an exciting color for special occasions.” Molly tucked the map into her coat and produced two party hats, one yellow, one blue, both decorated with suspiciously realistic fingerprints from previous operations. “Those are our uniforms. Mine says birthday, yours says helper, which is perfect because helpers are allowed near presents and rarely questioned about getaway bicycles.” Bill tried on the blue hat, which immediately slid over his eyes, transforming a master criminal into a very worried lampshade with excellent posture somehow. “Perhaps we should bring a gift after all. Something cheap, thoughtful, and light enough to carry while running from furious parents and their terriers outside.” From inside came a burst of laughter, followed by the unmistakable crash of a piñata meeting a lamp and losing decisively in front of witnesses everywhere. “New plan: we enter as emergency decorators, repair the lamp, accept gratitude, and leave carrying nothing except our dignity and possibly cupcakes for the road.” Bill looked at the invitation, then at Molly, whose expression had softened beneath the hat; even thieves could recognize a room full of delighted children. “Agreed. We steal one cupcake each, repair the lamp, and give the presents back to whoever brought the biggest box of socks with a receipt.”
created_at
"2026-09-23T08:27:28.282Z"
9/23/2026, 8:27:28 AM
model
"openai/gpt-6-luna-20260922"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
21255
moderation_latency
(null)
generation_time
33192
tokens_prompt
75
tokens_completion
935
native_tokens_prompt
66
native_tokens_completion
3684
native_tokens_completion_images
0
native_tokens_reasoning
3068
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
0.0018486
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.20.0; linux; x64))"
http_referer
(null)
request_id
"req-1790152048-oGRMWJ24VFBDJdCPBFJH"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790152048-IeFXvQIXKQilemVHgvg4"
upstream_id
"resp_0e7729d4a9662aaf016ab38d70a7fc87d1bb4ee7928b743bd7"
provider_responses
0
endpoint_id
"05e94e02-b9c2-4bbb-ba55-4082ee9ad687"
id
"resp_0e7729d4a9662aaf016ab38d70a7fc87d1bb4ee7928b743bd7"
is_byok
false
latency
1045
model_permaslug
"openai/gpt-6-luna-20260922"
provider_name
"OpenAI"
status
200
total_cost
0.0018486
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
100.00%
Matches word count
n/a
neededClean
false
words
500
100.00%
Dialogue to Total Word Ratio
Ratio: 50.00%, Deviation: 0.00%
neededClean
false
wordsTotal
500
wordsDialogue
250
100.0000%