NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-23-01M36P4T5YRSSZDSNXNERR3PYF
Run Details
Model:
GPT-6 Luna (Reasoning, Medium)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
1m
Cost:
$0.003072
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
27.17%
Matches word count
(details)
99.14%
Dialogue to Total Word Ratio
(details)
63.1543%
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
Rain tapped the bakery awning while Molly spread a hand-drawn map across a crate, its corners weighted by two muffins and Bill’s optimism about tomorrow. “The party starts at six. Presents sit by the cake, and nobody suspects two professionals. Also, Bill, your pencil is wearing a tutu for luck.” Bill studied the map upside down, then rotated it with grave ceremony, as if geography had personally insulted him once at school and never recovered. “I propose we enter disguised as entertainers. Children adore entertainers, and adults rarely question a man carrying a balloon shaped like a turnip.” Molly eyed his balloon, which sagged between them like a vegetable surrendering to gravity. Somewhere nearby, a kazoo rehearsed one mournful note again. “No turnip. We need a respectable cover. Perhaps we’re visiting cousins, bearing gifts, and leaving with identical gifts by accident, naturally, obviously.” She drew two stick figures beside the party tent, then gave one a cape and the other a suspiciously large mustache for reasons unexplained. “I can do mustache. My face is already halfway there. The other half is hidden by this scarf, which smells faintly of soup today.” A burst of children’s laughter floated over the fence. Molly’s expression softened, then tightened when she remembered their rent and Bill’s unpaid accordion lessons again. “We take only the presents. No cake, no toys, no tiny shoes. We have standards, and a conscience on alternate Thursdays, maybe even often.” Molly stared at him. Bill had once returned a library book during a fire drill, but had also stolen three traffic cones at noon. “Those cones were abandoned. And one followed me home. Anyway, the presents are practically asking for adventure in their colorful paper pajamas.” A paper crown skittered down the alley and landed on Bill’s head. It read BIRTHDAY KING, improving his posture considerably for almost three seconds. “Your Majesty, perhaps the king should reconsider. These children might notice their gifts missing and feel sad before the candles cool and ask questions loudly.” Bill removed the crown. His grin faded. From the tent came a chorus demanding the birthday child open the enormous box in silver stars. “We could leave the presents and steal the cake. Wait, no: that is somehow worse, especially if it has raisins, as Molly’s face clearly suggests.” Molly folded the map into a paper airplane and sent it toward the party. It landed in a tub of lemonade with a plop. “There, our plan has been washed away, which is more than can be said for the birthday boy’s shoes just now, anyway, in this alley.” They turned toward the bakery, where muffins waited. Behind them, the children cheered as the birthday child tore open a gift with three shrieks. “Good. They can keep their presents. We’ll celebrate our narrow escape with pastries and a firm promise to become better people together, starting Monday.”
created_at
"2026-09-23T08:29:33.764Z"
9/23/2026, 8:29:33 AM
model
"openai/gpt-6-luna-20260922"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
4012
moderation_latency
(null)
generation_time
60075
tokens_prompt
75
tokens_completion
1247
native_tokens_prompt
66
native_tokens_completion
6130
native_tokens_completion_images
0
native_tokens_reasoning
5527
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
0.0030716
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.20.0; linux; x64))"
http_referer
(null)
request_id
"req-1790152173-IDv6NfxfKtuP7YCov8XE"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790152173-6YME8dLemU85ek0ONs5i"
upstream_id
"resp_0da0eba6c5a6c411016ab38deddb8c87d1aa18c792b2d9453b"
provider_responses
0
endpoint_id
"05e94e02-b9c2-4bbb-ba55-4082ee9ad687"
id
"resp_0da0eba6c5a6c411016ab38deddb8c87d1aa18c792b2d9453b"
is_byok
false
latency
897
model_permaslug
"openai/gpt-6-luna-20260922"
provider_name
"OpenAI"
status
200
total_cost
0.0030716
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
27.17%
Matches word count
n/a
neededClean
false
words
481
99.14%
Dialogue to Total Word Ratio
Ratio: 49.59%, Deviation: 0.41%
neededClean
false
wordsTotal
482
wordsDialogue
239
63.1543%