NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-23-01M36RC6D0PJDDDT0KFP039J6A
Run Details
Model:
GPT-6 Luna (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
3m 3s
Cost:
$0.008871
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
77.27%
Dialogue to Total Word Ratio
(details)
88.6354%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 30% 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 parked the getaway van beneath a banner that read HAPPY BIRTHDAY, JUNIPER! The van was painted like a cupcake because Bill had insisted camouflage should be festive. Through the community hall’s open windows, they saw balloons, paper crowns, and a table sagging beneath gifts. Molly checked her watch twice. Bill adjusted a bow tie over his black sweater. The tie had tiny rubber ducks on it, which made him look less like a burglar and more like a nervous magician. “Does the tie inspire confidence?” Bill asked. “Only in ducks,” Molly said. “Remember, we’re taking presents, not children, cake, or green jelly.” “I made a list. Presents; cake only if emotionally available.” “Why bring a sack marked laundry?” “Loot attracts questions.” “What if they ask who we are?” “We’re the party’s accountants.” “That little girl’s watching. Smile like you know what a piñata is.” “It’s a papier-mâché tax shelter. Swing at it.” “Plan?” “Tell the truth and ask if they need help.” Bill reached for the handle, but it swung open. A boy in a paper crown held a clipboard, surveying them like a tiny customs officer. “Are you the entertainers?” he asked. Molly glanced at Bill’s duck tie. “We’re the surprise accountants.” “Good. The entertainer’s stuck in traffic. You’re in charge of the treasure hunt.” The boy crowned Molly and led them right inside. Frosting and orange juice scented the hall. A dozen children watched him point to gifts. “Hide these,” he instructed. “Then we’ll find them after cake.” Molly and Bill regarded the presents: bright ribbons, handmade tags. One read, “For Juniper, from Grandma. I knitted this badly but with love.” Another was a lumpy box decorated with a dog in sunglasses. Bill whispered, “The loot has become sentimental.” Molly whispered back, “Sentiment is just guilt wearing a nice hat.” They carried gifts to the stage, where Molly announced a treasure hunt with the confidence of someone who had never organized anything except an escape. Bill hid each present in plain sight: behind a balloon, beneath a chair, inside the cardboard castle. The children shrieked whenever they found one. Juniper opened the dog box and hugged a knitted scarf. “Grandma made it!” Bill watched her twirl, then quietly removed the laundry sack from his shoulder. He used it to gather stray wrapping paper instead. When the grown-ups brought out cake, Molly and Bill stood in line with the children. A frosting rose slid off Bill’s slice and landed on his duck tie. “Best job we’ve ever had,” he said. “We weren’t paid,” Molly replied. “We were paid in cake.” Juniper handed them each a party bag. Inside were stickers, a whistle, and two chocolate coins. Molly pocketed the coins. Bill put on the whistle, tried to look professional, and made a sound like a kettle having a crisis. At the van, Molly quietly placed the coins on the dashboard. “Next job,” Bill said, “we target a bank.” “Fine,” Molly said. “But only if there’s a birthday banner.”
created_at
"2026-09-23T09:08:32.815Z"
9/23/2026, 9:08:32 AM
model
"openai/gpt-6-luna-20260922"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
6732
moderation_latency
(null)
generation_time
183157
tokens_prompt
75
tokens_completion
3225
native_tokens_prompt
66
native_tokens_completion
17728
native_tokens_completion_images
0
native_tokens_reasoning
17062
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
0.0088706
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1790154512-YPTm6w6hhUhBFtvUvQDt"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790154512-maA67hYC8LazUujF9ie0"
upstream_id
"resp_00b0bec4145171f0016ab39710f0dc87d1869fbe2b3391b469"
provider_responses
0
endpoint_id
"05e94e02-b9c2-4bbb-ba55-4082ee9ad687"
id
"resp_00b0bec4145171f0016ab39710f0dc87d1869fbe2b3391b469"
is_byok
false
latency
1003
model_permaslug
"openai/gpt-6-luna-20260922"
provider_name
"OpenAI"
status
200
total_cost
0.0088706
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
77.27%
Dialogue to Total Word Ratio
Ratio: 32.27%, Deviation: 2.27%
neededClean
false
wordsTotal
502
wordsDialogue
162
88.6354%