NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-23-01M36R6S3BP2PW9F9AZE104XSN
Run Details
Model:
GPT-6 Luna (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
20.8s
Cost:
$0.000740
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
66.16%
Dialogue to Total Word Ratio
(details)
33.0799%
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 spread a crayon map across the diner table. It showed the community hall, the cake table, and a large, ominous cloud labeled CHILDREN. Beside it sat Bill, wearing a black turtleneck and the expression of a man who had recently lost an argument with a balloon. “Our plan is simple,” Molly said. “We enter the party, locate the presents, and leave with every last one.” “Simple,” Bill echoed. “I like simple. Does simple involve the bouncy castle?” “Only if it becomes necessary.” Bill eyed the map. “What’s the cloud?” “Children.” “Right. I was hoping it was weather.” Molly tapped the paper. She had drawn herself and Bill as stick figures wearing disguises: she had given herself a mustache; Bill’s disguise was a second mustache, placed directly beneath the first. The resemblance to a startled caterpillar was, she felt, convincing. The hall was hosting a birthday party for a seven-year-old named Ollie. The invitation had been pinned to a noticeboard outside their rented room, where it had announced a magician, a pony, and “presents welcome.” Molly had read that last phrase as a professional opportunity. Bill had read it as a warning. “I’m not sure we should rob children,” he said. “We’re not robbing children,” Molly replied. “We’re robbing a room in which children happen to be.” “That’s the sort of distinction that gets you chased by a pony.” She lowered her voice. “We’ll blend in. I’ll pose as an aunt. You’ll be the uncle.” “Whose uncle?” “Everyone’s, briefly.” Bill considered this. “And if someone asks which child I belong to?” “Look mysterious.” “I can do that. I once spent an entire bus ride pretending I knew where I was going.” Molly rolled up the map. “We wait for the magician’s act. While everyone watches him, we collect the gifts.” “What if the gifts are heavy?” “Then we’ll be grateful they’re generous.” The next afternoon, they arrived at the hall carrying a gift-wrapped box. It contained a brick, because Molly believed every convincing disguise required commitment. Bill wore a paper party hat and had attached a name tag reading UNCLE. Inside, children ran in joyful circles. A pony stood near the refreshments, looking like it had heard every bad plan in town and disapproved of all of them. Molly and Bill edged toward the presents, but a little girl intercepted them. “Are you Ollie’s aunt and uncle?” she asked. “Absolutely,” said Molly. “Which side?” “Both,” said Bill. “We’re very close.” The girl nodded, apparently satisfied, and handed them two cupcakes. Molly accepted hers. Bill accepted his with the solemnity of a diplomat receiving a treaty. Then the magician announced that his next trick required two volunteers. The children pointed at Molly and Bill. “We should go,” Molly whispered. “Can’t,” Bill whispered back. “I’ve been chosen by the birthday community.” Onstage, the magician tucked them into a velvet cabinet. When the doors opened, Molly and Bill emerged wearing glittery capes, while the children roared with approval. The pony snorted. Bill’s mustaches slid off and landed in his cupcake. At the gift table, Molly glanced at the colorful pile. Then Ollie ran over and hugged her around the waist. “Thanks for coming, Auntie!” Molly looked at Bill. He looked at the presents, then at the boy. “Maybe,” Bill murmured, “we leave the presents where they are.” Molly sighed. “Fine. But we’re taking the cupcakes.” They left with frosting on their sleeves, a brick still wrapped, and the pony’s grudging respect.
created_at
"2026-09-23T09:05:35.353Z"
9/23/2026, 9:05:35 AM
model
"openai/gpt-6-luna-20260922"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
6498
moderation_latency
(null)
generation_time
20783
tokens_prompt
75
tokens_completion
1222
native_tokens_prompt
66
native_tokens_completion
1466
native_tokens_completion_images
0
native_tokens_reasoning
679
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
0.0007396
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1790154335-xIOK5gI5OjBBuGHUZDl1"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790154335-8KoAm5bHCpeJoMLnP9tR"
upstream_id
"resp_04866c84570906a2016ab3965f741887d18b5ef4558fc01d85"
provider_responses
0
endpoint_id
"05e94e02-b9c2-4bbb-ba55-4082ee9ad687"
id
"resp_04866c84570906a2016ab3965f741887d18b5ef4558fc01d85"
is_byok
false
latency
397
model_permaslug
"openai/gpt-6-luna-20260922"
provider_name
"OpenAI"
status
200
total_cost
0.0007396
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
578
66.16%
Dialogue to Total Word Ratio
Ratio: 32.87%, Deviation: 2.87%
neededClean
false
wordsTotal
581
wordsDialogue
191
33.0799%