NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-23-01M37HD9QKEYH16D0H453NPNBX
Run Details
Model:
GPT-6 Sol (Reasoning, Medium)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
27.8s
Cost:
$0.013812
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
0.01%
Matches word count
(details)
51.47%
Dialogue to Total Word Ratio
(details)
25.7378%
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
Molly spread a crayon drawing across the café table. It showed a house, a mountain of presents, and a stick figure labeled BILL, whose arms appeared to have been drawn by someone with a grudge. Outside, rain polished the windows. Inside, Bill inspected the drawing upside down. “Why am I shaped like a coat hanger? And why have you drawn yourself wearing a crown? We agreed this was a partnership.” “The crown is a disguise. Princess Molly can stroll into a children’s party without suspicion. Coat Hanger Bill has to explain himself.” Bill turned the page around. Saturday’s birthday party would fill the community hall with children, parents, cake, and, according to Molly’s cousin, at least thirty wrapped presents. He had suggested stealing one. Molly had called that a lack of ambition. “I could be a magician,” he said. “Children love magicians. I’ll pull a rabbit from a hat while you collect the presents.” “Can you do that? Because last time you tried, you pulled out a second hat and frightened yourself half to death.” Molly had borrowed a party invitation from her cousin’s refrigerator. She tapped the gold-foil heading: DETECTIVE DAY. Beneath it, a smiling cartoon magnifying glass invited guests to solve mysteries. Bill studied it with the solemnity of a man reviewing his own arrest warrant. “That’s unfortunate. We’ll have detectives everywhere. Small detectives, granted, but they’ve got energy, and they can fit under tables.” “They’ll be busy finding clues. We walk in as caterers, stack the gifts in cake boxes, and leave before anyone notices.” Bill glanced at the invitation’s back. It listed games, allergies, and a warning that the birthday child disliked balloons. Molly had underlined “cake served at two” three times, as though it were the weak point in a fortress. “What if somebody asks what kind of caterers we are? You know nothing about food, and I once burned a cucumber.” “We say we’re dessert specialists. We look professionally disappointed in the cake, and nobody will dare ask us another question.” A paper bag sat beside Molly’s chair. She opened it to reveal two white aprons and a packet of adhesive mustaches. Bill lifted one, held it beneath his nose, and immediately sneezed it onto the drawing. “Why do caterers need mustaches? Also, why does this apron say ‘Junior Detective’ in enormous letters across the front?” Molly snatched the apron. Tucked inside was a note in purple felt-tip, addressed to them by name. She read it twice. Bill watched her expression change from outrage to the wary respect she reserved for excellent locks. “The birthday child knows our plan,” she said. “She says we can play the criminals if we bring our own disguises.” “Does she mention payment? I’m not getting chased around a hall by seven-year-olds for nothing, especially if that cake’s disappointing.” Molly checked the note again. It promised two juice boxes and first choice of cupcakes. Beneath that, in smaller writing, was a request that they arrive ten minutes early to help set up the presents. She folded the note and pocketed it. “First choice is worth something,” she said. “But we’re not carrying thirty presents in for free. I’m negotiating sprinkles.”
created_at
"2026-09-23T16:26:03.39Z"
9/23/2026, 4:26:03 PM
model
"openai/gpt-6-sol-20260922"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
9332
moderation_latency
(null)
generation_time
27720
tokens_prompt
75
tokens_completion
960
native_tokens_prompt
66
native_tokens_completion
1368
native_tokens_completion_images
0
native_tokens_reasoning
682
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
0.013812
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.20.0; linux; x64))"
http_referer
(null)
request_id
"req-1790180763-b3Nh94qPLle96w9ayOWF"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790180763-C7QRFbWSGQWRvqZEmqZ0"
upstream_id
"resp_0f6825b120188dd1016ab3fd9b910c87d19f85c57283ae9955"
provider_responses
0
endpoint_id
"1cfc7d9d-4404-4b8e-9ee4-58ae45c9dcd4"
id
"resp_0f6825b120188dd1016ab3fd9b910c87d19f85c57283ae9955"
is_byok
false
latency
639
model_permaslug
"openai/gpt-6-sol-20260922"
provider_name
"OpenAI"
status
200
total_cost
0.013812
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
0.01%
Matches word count
n/a
neededClean
false
words
531
51.47%
Dialogue to Total Word Ratio
Ratio: 46.36%, Deviation: 3.64%
neededClean
false
wordsTotal
535
wordsDialogue
248
25.7378%