NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-09-01KX40VJ5XQMF1HZ6FKFMKMP56
Run Details
Model:
GPT-5.6 Sol (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
59.5s
Cost:
$0.092760
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
100.00%
Dialogue to Total Word Ratio
(details)
99.9990%
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 crayon map across the café table, pinning its corners with sugar bowls. Bill squinted at it, upside down and suspiciously close to the jam beside his elbow. “The target is the Buttercup Community Hall at three o’clock. Forty children, eighty adults, one magician, and a mountain of presents beneath a banner reading Happy Birthday, Poppy. We enter disguised as entertainers, locate the gift table, and remove everything during the chaos of musical statues. Simple, elegant, festive, criminal.” “I see two problems. First, neither of us can entertain children. Second, musical statues has pauses, and I panic whenever music stops. At my cousin’s wedding I froze beside the cake for forty minutes. People put coats on me and called me the rack. It damaged my professional confidence considerably.” Bill tapped the map, leaving a marmalade fingerprint over the emergency exit. Molly rotated the paper and revealed he had been studying a menu for waffles. The map lay underneath. “You won’t entertain. You’ll be Professor Pickle, balloon scientist. Wear the green coat, carry long balloons, and say science whenever one squeaks. Children respect authority, especially authority holding inflatable sausages. I’ll be Madame Marzipan, face painter and tactical gift relocation specialist. I have brushes, glitter, and absolutely no artistic restraint.” “You painted a tiger on me once. It looked like an anxious potato. Also, how do we carry forty presents out? The red wagon squeaks, the blue van smells of haddock, and your bicycle basket barely holds a sandwich without emotional collapse. We need transport with dignity and adequate suspension.” Outside, a bus sighed at the curb. Molly watched three parents wrestle a papier-mâché castle through the hall doors, then drew a small crown beside the kitchen with purple pencil. “The castle is our Trojan horse. We hide inside until the cake arrives, then roll the presents through the kitchen into the catering truck. I borrowed uniforms. We stack parcels in empty vegetable crates and label them sprouts. Nobody willingly investigates sprouts at a party. They are nature’s security system.” Molly leaned back. The chair squealed like a betrayed mouse. Bill flinched, knocking over a sugar bowl and catching it neatly in his hat. Neither considered this especially reassuring practice. “Borrowed means stolen, doesn’t it? And catering trucks have caterers. Caterers are observant. They count spoons. My aunt catered funerals and could detect a missing vol-au-vent across a crowded chapel. What if someone opens a crate and finds a train set instead of sprouts? We claim it is seasonal produce.” Molly took the sugar bowl from his hat and slid him a photograph of Poppy beside a tower of wrapped boxes. Bill studied it solemnly. One parcel resembled France exactly. “Relax. Poppy is my niece. She invited me last week, but her mother said absolutely no stealing presents. So we steal them, hide them in the truck, then return them dramatically when the magician fails. We become heroes, and keep the catering uniforms. Plus any cake abandoned during the investigation.”
created_at
"2026-07-09T18:04:50.758Z"
7/9/2026, 6:04:50 PM
model
"openai/gpt-5.6-sol-20260709"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
34755
moderation_latency
(null)
generation_time
59403
tokens_prompt
77
tokens_completion
820
native_tokens_prompt
66
native_tokens_completion
3081
native_tokens_completion_images
0
native_tokens_reasoning
2441
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
0.09276
router
(null)
provider_responses
0
endpoint_id
"a54c5de0-89bf-4ad7-a212-cf977eed918a"
id
"resp_0408ecb761e9992d016a4fe2c2da5881968034c0ec055d839c"
is_byok
false
latency
270
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-1783620290-yYhtGvAiO0XKeP0F2aJO"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1783620290-065XNbsCkrNSEEHaEuKU"
upstream_id
"resp_0408ecb761e9992d016a4fe2c2da5881968034c0ec055d839c"
total_cost
0.09276
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
100.00%
Matches word count
n/a
neededClean
false
words
500
100.00%
Dialogue to Total Word Ratio
Ratio: 69.98%, Deviation: 0.02%
neededClean
false
wordsTotal
503
wordsDialogue
352
99.9990%