NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-23-01M36R7VBSYBJXFTAMATX7R1AA
Run Details
Model:
GPT-6 Luna (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
16.8s
Cost:
$0.000638
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.06%
Dialogue to Total Word Ratio
(details)
0.0297%
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
“The party starts at three,” Molly said, spreading a hand-drawn map across the diner table. “At four, the children are distracted by cake. At five, the presents are unattended.” Bill peered at the map. “This is a placemat.” “It has the building on it.” “It has a smiling pancake on it.” “That’s the community hall.” “The community hall has syrup?” Molly folded the placemat in half. “We’re breaking in, collecting every present, and leaving before anyone notices.” “Before anyone notices what?” “That the presents are gone.” “I feel like the children may notice.” “Children notice everything. They once spotted me hiding a cupcake behind my back.” “Was it because the cupcake was behind your back?” “It was because I was eating it.” Bill glanced toward the window, where a paper banner across the street read HAPPY BIRTHDAY, TREVOR! “I’m not sure I can rob a birthday party.” “You can. You’re just letting the balloons intimidate you.” “One of them says ‘You’re a big kid now.’ It knows my weaknesses.” “We go in, take the gifts, go out. Simple.” “Your last simple plan ended with me dressed as a garden gnome at a retirement home.” “You were convincing.” “They tried to water me.” Molly tapped the placemat. “Focus. What do children want most?” “Cake?” “Besides cake.” “More cake?” “Fine. What do thieves want most?” “An alibi?” “Bill.” “Presents, I suppose.” “Exactly. And we take all of them.” He stared at the banner again. “There are a lot of presents.” “How many?” “I can see twelve from here.” “You’re counting the ones in the window?” “I’m counting the ones stacked to the ceiling in my imagination.” Molly lowered her voice. “We can’t leave any behind. That’s how you get caught. You take one, they ask questions. You take all of them, they assume the party was a dream.” “That is not how dreams work.” “You don’t know. You’ve never had a birthday party.” “I had one.” “You invited the mailman?” “He brought a package. I thought it was a social call.” Molly sighed. “Look, we’ll be in and out before the candles are lit.” “Suppose a child asks what we’re doing?” “We say we’re the present inspectors.” “Do present inspectors exist?” “Not if we do this properly.” Bill pushed the placemat away. “Molly, those gifts are for a kid. He’s turning six.” “Six is old enough to understand disappointment.” “Is it?” “I understood it at six.” “Because your uncle gave you a sweater?” “Because my uncle gave me a sweater with a picture of himself on it.” Bill looked toward the party hall. A little boy pressed his face to the glass, holding up a crooked paper crown. Molly followed his gaze. “Don’t look at him.” “I wasn’t.” “He’s waving.” “I thought he was signaling for help.” “He’s waving, Bill.” Bill stood. “I’m going over there.” “To steal the presents?” “To ask if they need help carrying them inside.” Molly rose too, snatching up the placemat. “That is not the plan.” “It is now. And afterward, we can steal the cake.” She considered this. “Only if it’s chocolate.” “It’s a children’s party. Of course it’s chocolate.” “Then move, present inspector.” “I thought we weren’t inventing that job.” “We are now.”
created_at
"2026-09-23T09:06:10.436Z"
9/23/2026, 9:06:10 AM
model
"openai/gpt-6-luna-20260922"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
8195
moderation_latency
(null)
generation_time
16731
tokens_prompt
75
tokens_completion
1074
native_tokens_prompt
66
native_tokens_completion
1263
native_tokens_completion_images
0
native_tokens_reasoning
516
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
0.0006381
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1790154370-KdHY1eK2qeEC21V1GrWQ"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790154370-FTbIpwhJ8xFu4gfddwZX"
upstream_id
"resp_05339cf2d33f6a07016ab396828e2487d1a8c349919d322010"
provider_responses
0
endpoint_id
"05e94e02-b9c2-4bbb-ba55-4082ee9ad687"
id
"resp_05339cf2d33f6a07016ab396828e2487d1a8c349919d322010"
is_byok
false
latency
406
model_permaslug
"openai/gpt-6-luna-20260922"
provider_name
"OpenAI"
status
200
total_cost
0.0006381
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
538
0.06%
Dialogue to Total Word Ratio
Ratio: 82.19%, Deviation: 12.19%
neededClean
false
wordsTotal
539
wordsDialogue
443
0.0297%