NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-23-01M37HV9CN4M3EQE837C0188B9
Run Details
Model:
GPT-6 Sol (Reasoning, Medium)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
19.9s
Cost:
$0.009832
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
43.38%
Matches word count
(details)
1.14%
Dialogue to Total Word Ratio
(details)
22.2592%
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
The invitation lay between them on the café table, glittering beneath a smear of ketchup. Molly had obtained it from a recycling bin. Bill had brought binoculars, a balloon pump, and the haunted expression of a man who had once lost an argument with a piñata. Outside, rain blurred the windows. “Seven years old,” Molly said. “Twenty guests. Twenty presents, minimum. I counted three grandparents on the guest list, so we might be looking at thirty.” “Your maths assumes the grandparents understand boundaries,” Bill said. “My nan once gave me a live goose. I was thirty-four and allergic to feathers.” Molly turned the invitation over. A cartoon dinosaur smiled above the words MOSTLY INDOORS. She had hoped for an outdoor party. Outdoors meant plenty of space and adults whose attention could be diverted by weather. Indoors meant a hallway full of tiny shoes and somebody’s father standing beside the coats, asking questions. “We blend in,” she said. “You’re a magician. I’m your assistant.” “You can’t be my assistant. You’re better at magic than I am.” “Exactly. I do the tricks. You wear a cape and make the children expect less.” Bill considered this. Across the café, a toddler dropped a spoon and looked delighted with the result. Bill looked away first. He had faced museum guards, furious dogs, and one surprisingly athletic mayor, but children unsettled him. They noticed everything, particularly the things adults had paid good money to conceal. “What if they ask for a rabbit?” he said. “We tell them the rabbit is on holiday.” “Seven-year-olds don’t believe rabbits take holidays.” “Then you’ve underestimated rabbits. Pack the cape.” Molly drew a rough rectangle on a napkin to represent the party room. Before she could add a doorway, a woman at the next table leaned over. She had been listening for some time, judging by her smile. Molly covered the napkin with her hand, though the drawing looked more like a biscuit. “Sorry,” the woman said. “Are you two the entertainers for Alfie’s party? I’m his mother.” Bill made a sound like a balloon surrendering. “Yes,” Molly said. “And we’re terribly excited.” “Wonderful. One small change: the party’s at the hospital now. Alfie’s little sister can’t come home yet, and he wouldn’t celebrate without her.” The woman returned to her coffee. Molly and Bill sat quietly. Rain tapped at the window. Bill folded his binoculars into their case; Molly smoothed the glittering invitation flat. Across the café, the toddler dropped the spoon again. Neither thief offered to pick it up, but both watched until someone did. “We can’t steal presents from a hospital,” Bill whispered. “No,” Molly said. “We’d have to steal them before they got there.” “Molly.” “I’m joking. Put the balloon pump away.” Bill slipped it under the table. “We could still go,” he said. “Do a show for free.” “You don’t know any tricks.” “I can make a goose disappear. Eventually.” Molly began to laugh, then stopped to consider the balloon pump. “Can you make balloon animals?” “No. But I can make balloon questions.” “Good enough,” she said, and picked up the invitation.
created_at
"2026-09-23T16:33:41.788Z"
9/23/2026, 4:33:41 PM
model
"openai/gpt-6-sol-20260922"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
7998
moderation_latency
(null)
generation_time
19837
tokens_prompt
75
tokens_completion
855
native_tokens_prompt
66
native_tokens_completion
970
native_tokens_completion_images
0
native_tokens_reasoning
287
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
0.009832
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.20.0; linux; x64))"
http_referer
(null)
request_id
"req-1790181221-uNIJ71GbtTXVBCyAaxzS"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790181221-EsBu8SJM976vQiCubU1M"
upstream_id
"resp_05d49f44519d285b016ab3ff65e50887d1a3fadc96e965c143"
provider_responses
0
endpoint_id
"1cfc7d9d-4404-4b8e-9ee4-58ae45c9dcd4"
id
"resp_05d49f44519d285b016ab3ff65e50887d1a3fadc96e965c143"
is_byok
false
latency
1123
model_permaslug
"openai/gpt-6-sol-20260922"
provider_name
"OpenAI"
status
200
total_cost
0.009832
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
43.38%
Matches word count
n/a
neededClean
false
words
517
1.14%
Dialogue to Total Word Ratio
Ratio: 40.54%, Deviation: 9.46%
neededClean
false
wordsTotal
518
wordsDialogue
210
22.2592%