NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-23-01M36P0GYCAB1SF8ZSXN003R26
Run Details
Model:
GPT-6 Luna (Reasoning, Medium)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
12.6s
Cost:
$0.000524
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
95.99%
Matches word count
(details)
59.39%
Dialogue to Total Word Ratio
(details)
77.6864%
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
“Tonight,” Molly said, spreading a crayon drawing across the diner table, “we take every present from the children’s party.” Bill studied the drawing. “This is a picture of a cake.” “It’s a map.” “The cake has a dinosaur on it.” “That’s the security system.” A waitress passed and glanced at their solemn faces. Bill folded the drawing over the dinosaur. “Fine,” he whispered. “What’s our plan?” “We enter as entertainers.” “You can’t juggle.” “I can juggle one orange.” “That’s not entertainment. That’s lunch.” “I’ll be the magician. You’ll be my assistant.” “I’m not wearing a sequined cape again.” “You said the cape made you feel powerful.” “It made me feel like a sofa at a casino.” Molly tapped the drawing. “We perform a dazzling show, everyone watches us, and we collect the presents.” “Collect them where?” “Somewhere discreet.” “Like a closet?” “A closet is where people keep things. That’s practically permission.” “It is not.” “Then we need a better plan.” Outside, a bus rolled by with a cheerful picture of a smiling clown painted on its side. Bill looked at it, then at Molly. “Could we pretend to be clowns?” “Bill, children can spot a fake clown.” “Can adults?” “Adults are worse. They’ll ask us to twist balloons into animals.” “I can make a balloon animal.” “What?” “A balloon.” Molly sighed. “We need costumes, a distraction, and a getaway.” “Those are three separate plans.” “That’s why we’re professionals.” “You stole a garden gnome last week and returned it because it looked lonely.” “It was staring at me.” “It’s a gnome.” “Exactly. No one would believe it had those feelings unless they were real.” Bill pinched the bridge of his nose. “Molly, these are children’s presents. There could be toys, games, maybe one of those little plastic dinosaurs.” “Don’t tempt me with dinosaurs.” “I’m saying the children will miss them.” “They’ll have cake.” “Cake isn’t a replacement for a present.” “It can be, if you put a bow on it.” “That’s not how gifts work.” Molly stared at the crayon map. The dinosaur’s paper eyes seemed to stare back. “What if we take the presents, then bring them back after the party?” “That’s still stealing.” “Borrowing with dramatic timing.” “Also stealing.” A silence settled between them, interrupted by the waitress placing two slices of pie on the table. “Did either of you order these?” she asked. “No,” said Bill. Molly eyed the pie. “Perhaps the universe is rewarding our honesty.” “It’s probably a mistake.” “Then we should return it.” Bill pushed his plate away. “Look at us. Already thwarting a pastry heist.” Molly stood, folded the drawing, and tucked it into her coat. “New plan. We go to the party, perform our act, and leave the presents exactly where they are.” “What’s the act?” “You juggle an orange.” “And you?” “I’ll tell the dinosaur a joke.” “That’s the security system?” “No,” Molly said, opening the diner door. “That’s the guest of honor.” Bill followed her out. “I hope it likes pie.” “Everybody likes pie,” Molly said. “Except the gnome.”
created_at
"2026-09-23T08:27:13.236Z"
9/23/2026, 8:27:13 AM
model
"openai/gpt-6-luna-20260922"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
5266
moderation_latency
(null)
generation_time
12551
tokens_prompt
75
tokens_completion
919
native_tokens_prompt
66
native_tokens_completion
1034
native_tokens_completion_images
0
native_tokens_reasoning
326
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
0.0005236
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.20.0; linux; x64))"
http_referer
(null)
request_id
"req-1790152033-wJqNF5sP4ZxmM1HBZmPV"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790152033-DxXbNm7FlgEuvHoiUhRF"
upstream_id
"resp_05c25cd21c26e39b016ab38d615b1887d1988e39377da5bb4f"
provider_responses
0
endpoint_id
"05e94e02-b9c2-4bbb-ba55-4082ee9ad687"
id
"resp_05c25cd21c26e39b016ab38d615b1887d1988e39377da5bb4f"
is_byok
false
latency
627
model_permaslug
"openai/gpt-6-luna-20260922"
provider_name
"OpenAI"
status
200
total_cost
0.0005236
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
95.99%
Matches word count
n/a
neededClean
false
words
508
59.39%
Dialogue to Total Word Ratio
Ratio: 73.23%, Deviation: 3.23%
neededClean
false
wordsTotal
508
wordsDialogue
372
77.6864%