NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-23-01M36RDK1FJ70RXPABNMCFAW9A
Run Details
Model:
GPT-6 Luna (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
22.6s
Cost:
$0.000816
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
7.22%
Dialogue to Total Word Ratio
(details)
53.6076%
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
Molly and Bill crouched behind a hedge, watching a garden full of balloons bob in the evening breeze. A banner read HAPPY BIRTHDAY, THEO. Beneath it stood a mountain of wrapped presents. “Remember,” Molly whispered, “we’re here to steal every present.” “I remember. I’m trying to forget the clown.” “The clown is harmless.” “He has a shoe phone.” “It’s a normal shoe.” “It rang, Molly.” “Maybe he has a very small foot.” Bill peered through the leaves. “There are at least twenty gifts.” “Excellent.” “Some are enormous.” “More excellent.” “One has a ribbon the size of a washing line.” “Don’t get distracted by the ribbon.” “I’m not distracted. I’m estimating how many of us it would take to carry it.” “There are two of us.” “Then we’ll need to make two trips.” “We are not making two trips. We are making one smooth, elegant exit.” “Should I bring the getaway wagon?” “You brought a wagon?” “I brought a wheelbarrow.” “Bill.” “It has a little bell.” Molly rubbed her forehead. Beyond the hedge, children chased bubbles while a parent handed out paper plates. The clown waved at them. Bill waved back before Molly could stop him. “Why did you wave?” she hissed. “He looked lonely.” “He’s a clown at a birthday party.” “Exactly. He works in a difficult field.” Molly studied the presents. “We wait until everyone’s distracted, then we—” “Sing?” “What?” “The children are singing. We could join in. Blend with the crowd.” “You can’t sing.” “I can sing one note.” “You can hum one note.” “Then I’ll hum convincingly.” “The plan is not to become part of the party.” “Why not? They have cake.” “We’re thieves.” “Thieves can enjoy cake.” A child appeared beside the hedge, holding a cupcake with one candle. “Are you hiding?” Bill looked at Molly. Molly looked at Bill. “No,” Molly said. “Yes,” said Bill. The child nodded, apparently satisfied. “You can come in. Theo says everyone gets cake.” Molly’s expression softened despite herself. “Everyone?” “Even the clown,” said the child. “He gets two pieces because he’s sad.” Bill stood. “That’s a generous policy.” “Bill,” Molly warned. “What? I’m gathering intelligence.” The child offered him the cupcake. Bill accepted it with the solemnity of a knight receiving a sword. Molly glanced again at the gifts. “We could still take them.” “We could,” Bill said. “But then we’d miss cake.” “And the children would lose their presents.” “And the clown might get sadder.” Molly sighed. “You’re making this very difficult.” “I’m making it delicious.” They stepped out from behind the hedge. The parent greeted them, mistaking them for late-arriving guests, and handed Molly a paper hat. Bill was immediately recruited to lead the birthday song. As everyone sang, he leaned toward Molly. “On the bright side, we’re now inside.” “On the brighter side,” she murmured, “we’re leaving with exactly what we came for.” “Cake?” “Cake.” “And the wheelbarrow?” “Leave it by the hedge.” Bill frowned. “But it has a bell.” “Then let it ring for the children.”
created_at
"2026-09-23T09:09:18.525Z"
9/23/2026, 9:09:18 AM
model
"openai/gpt-6-luna-20260922"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
8530
moderation_latency
(null)
generation_time
22522
tokens_prompt
75
tokens_completion
1017
native_tokens_prompt
66
native_tokens_completion
1619
native_tokens_completion_images
0
native_tokens_reasoning
911
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
0.0008161
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1790154558-uCSexW8JAwQl0gSpahcT"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790154558-a0OO3psG0noe1mx2GUxC"
upstream_id
"resp_093e239bb679b154016ab3973eadf487d1b6d8cf6af5cdc2a9"
provider_responses
0
endpoint_id
"05e94e02-b9c2-4bbb-ba55-4082ee9ad687"
id
"resp_093e239bb679b154016ab3973eadf487d1b6d8cf6af5cdc2a9"
is_byok
false
latency
748
model_permaslug
"openai/gpt-6-luna-20260922"
provider_name
"OpenAI"
status
200
total_cost
0.0008161
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
501
7.22%
Dialogue to Total Word Ratio
Ratio: 62.75%, Deviation: 7.25%
neededClean
false
wordsTotal
502
wordsDialogue
315
53.6076%