NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-23-01M37HARQRJCJJJ3CW3X5FXMA8
Run Details
Model:
GPT-6 Sol
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
48.1s
Cost:
$0.006912
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
68.10%
Matches word count
(details)
0.29%
Dialogue to Total Word Ratio
(details)
34.1952%
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 spread a floor plan across the café table. It was drawn in crayon and labeled THE CASTLE OF BIRTHDAY. “Your intelligence source is a child?” Bill asked. “My intelligence source is the child’s grandmother. This is what she had in her handbag.” “The grandmother drew it?” “She said the birthday girl did. See that purple blob? Kitchen.” “I thought it was a dragon.” “The dragon is in the kitchen. It’s a themed party.” Bill lowered his voice. “Remind me why we’re stealing presents from a seven-year-old.” “We’re not. We’re stealing them from the parents who accidentally had them delivered to the wrong address.” “All of them?” “We don’t know which one contains the emerald.” “So our plan is to steal every present at a children’s party.” “Temporarily. We unwrap them, find the emerald, and return the rest.” “That’s not a plan. That’s an apology with logistics.” Molly tapped the drawing. “Party starts at two. We arrive at two fifteen, dressed as entertainers.” “I’m not wearing the unicorn costume again.” “You wore it beautifully.” “A child tried to feed me a carrot through the eyehole.” “Because you stayed in character.” “Because I couldn’t find the zipper.” A waiter approached. Molly folded the map, revealing a menu beneath it. They ordered coffees and waited until he left. “Fine,” Bill said. “What am I wearing?” “Knight’s armor.” “That’s worse. I’ll sound like a drawer full of cutlery.” “Perfect. They’ve booked a princess party. Nobody will question a knight.” “They might question a knight carrying a sack of presents.” “You’ll carry a laundry bag.” “Oh, that’s much less suspicious.” “I’ll announce a surprise game. Everyone puts their gifts in the bag. Then I take them to the ‘royal vault’ for safekeeping.” “Where’s the vault?” “Guest bathroom.” Bill studied the crayon drawing. “That purple blob?” “Dragon. The bathroom is the green rectangle.” “Then why does it say POTTY DRAGON?” “Because the dragon has a name.” He leaned back. “We need a better exit. The front door faces the garden.” “Side gate?” “Locked.” “Fence?” “I’m wearing armor.” “Take it off.” “In the bathroom with forty stolen gifts?” “Borrow the dragon’s wings.” Their coffees arrived. Bill waited until the waiter was gone, then whispered, “What if the emerald isn’t there?” “Then we put everything back before cake.” “And if someone catches us?” “We say it’s pass-the-parcel.” “With every parcel?” “Birthday deluxe.” Bill stirred his coffee. “You know what worries me most? Not the parents. Not the police. The birthday girl.” “What about her?” “She drew a surveillance camera on the staircase.” Molly looked again. Beside a pink spiral sat a tiny black square. “That,” she said, “is a spider.” “It has an arrow pointing to it. The arrow says MUM CAM.” Molly folded the plan. “Right. Revised plan. We ask the parents for the delivery box.” “And the costumes?” “We wear them.” “Why?” “Because the princess booked them, Bill. We’re not monsters.”
created_at
"2026-09-23T16:24:40.449Z"
9/23/2026, 4:24:40 PM
model
"openai/gpt-6-sol-20260922"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
5729
moderation_latency
(null)
generation_time
48054
tokens_prompt
75
tokens_completion
888
native_tokens_prompt
66
native_tokens_completion
678
native_tokens_completion_images
0
native_tokens_reasoning
0
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
0.006912
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.20.0; linux; x64))"
http_referer
(null)
request_id
"req-1790180680-hCDcMBEKQGEqTAwM22Ma"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790180680-Q1Fr09cS4hc79hjxPMa7"
upstream_id
"resp_0308eed8afe84adf016ab3fd488db087d1801cd55acdefea8f"
provider_responses
0
endpoint_id
"1cfc7d9d-4404-4b8e-9ee4-58ae45c9dcd4"
id
"resp_0308eed8afe84adf016ab3fd488db087d1801cd55acdefea8f"
is_byok
false
latency
707
model_permaslug
"openai/gpt-6-sol-20260922"
provider_name
"OpenAI"
status
200
total_cost
0.006912
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
68.10%
Matches word count
n/a
neededClean
false
words
486
0.29%
Dialogue to Total Word Ratio
Ratio: 80.82%, Deviation: 10.82%
neededClean
false
wordsTotal
490
wordsDialogue
396
34.1952%