NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-09-01KX41A60ZNC385HZTZ0J16ZYJ
Run Details
Model:
GPT-5.6 Sol
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
16.0s
Cost:
$0.022560
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
60.28%
Matches word count
(details)
0.03%
Dialogue to Total Word Ratio
(details)
30.1504%
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 the crayon map across the café table. “The target is here.” Bill squinted. “That’s a drawing of a dragon eating a postman.” “It’s a floor plan.” “Then why is the kitchen breathing fire?” “Because I borrowed the map from the birthday boy. Focus. At three o’clock, twenty-six children enter Happy Hippo Party Palace. At three-ten, presents are stacked beside the throne.” “What throne?” “The birthday throne.” “You can’t just add ‘birthday’ to furniture and make it sinister.” “Birthday gallows.” “Stop.” “Birthday iron maiden.” “Molly.” “Fine. The presents remain unguarded while everyone watches Professor Pudding’s Magic Explosion.” Bill lowered his voice. “What explodes?” “Pudding, presumably.” “Presumably? I’m not risking custard shrapnel.” “You’ll be wearing protective gear.” “A paper crown is not protective gear.” “It’ll protect your identity.” “I’m forty-three. A paper crown makes me look like a bankrupt king.” “Exactly. Nobody suspects royalty.” A waiter approached. Molly folded the map. Bill smiled. “We’re planning a surprise party.” “For children?” asked the waiter. “For parole officers,” Molly said. The waiter retreated. Bill sighed. “How do we get inside?” “Disguises. I’ll be Princess Sparkleberry.” “You don’t look like Princess Sparkleberry.” “Nobody does. She’s fictional.” “And me?” “Gordon the Giraffe.” “I’m not wearing the neck.” “The neck is essential.” “The neck got stuck in a ceiling fan last time.” “That ceiling fan was unusually aggressive.” “It carried me through a conservatory.” “You made the local news.” “My mother recorded it.” “Publicity is useful.” “She plays it at Christmas.” Molly tapped the map. “We enter through the balloon arch, mingle, and wait for the magician. When the lights dim, you create a diversion.” “What kind?” “Release the ferret.” Bill stared. “We have a ferret?” “We will after lunch.” “I’m allergic.” “To lunch?” “To ferrets.” “Then don’t inhale it.” “And while everyone screams?” “I load the presents into the laundry cart.” “Why is there a laundry cart?” “Children generate jam.” “That sentence explains too much.” “We wheel the haul through the kitchen and escape in the ice-cream van.” “Do we own an ice-cream van?” “We will temporarily.” “That means steal.” “It means borrow without paperwork.” “What if the children chase us?” “They’re six.” “They’re fast, Molly. Low center of gravity. No pensions to lose.” “I’ve prepared obstacles.” “Please don’t say marbles.” “Cupcakes.” “You’ll waste cupcakes?” “For the operation.” Bill looked horrified. “We’re thieves, not monsters.” A balloon popped near the counter. Both ducked. Molly rose first. “Nerves?” “Professional reflexes.” “You hid behind a scone.” “It was structural cover.” She slid him a photograph. “Here’s the present table.” Bill studied the mountain of bright parcels. “Hold on. Those labels say ‘For Children’s Hospital.’” Molly grabbed the picture. “What?” “It’s a charity party.” Silence settled between them. “We can’t steal those,” Molly said. “Obviously.” “We’ll steal from somewhere else.” “Obviously.” Molly refolded the dragon map. “What about Professor Pudding?” Bill smiled slowly. “Magicians carry cash.” “And explosive custard.” “We’ll wear crowns.”
created_at
"2026-07-09T18:12:49.832Z"
7/9/2026, 6:12:49 PM
model
"openai/gpt-5.6-sol-20260709"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
722
moderation_latency
(null)
generation_time
15936
tokens_prompt
77
tokens_completion
1102
native_tokens_prompt
66
native_tokens_completion
741
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
0.02256
router
(null)
provider_responses
0
endpoint_id
"a54c5de0-89bf-4ad7-a212-cf977eed918a"
id
"resp_01197fa5c49b9042016a4fe4a1efb081938b08bf3194f8f0e3"
is_byok
false
latency
214
model_permaslug
"openai/gpt-5.6-sol-20260709"
provider_name
"OpenAI"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.18.0; linux; x64))"
http_referer
(null)
request_id
"req-1783620769-xNNtIb0jMu1N2CUKxwof"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1783620769-kalLgaB51235iuBCsjnn"
upstream_id
"resp_01197fa5c49b9042016a4fe4a1efb081938b08bf3194f8f0e3"
total_cost
0.02256
cache_discount
(null)
upstream_inference_cost
0
provider_name
"OpenAI"
response_cache_source_id
(null)
data_region
"global"
Evaluation details
Result
Evaluator
Details
Meta Data
60.28%
Matches word count
n/a
neededClean
false
words
485
0.03%
Dialogue to Total Word Ratio
Ratio: 82.86%, Deviation: 12.86%
neededClean
false
wordsTotal
490
wordsDialogue
406
30.1504%