NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-09-01KX40W5524EY2NMWJ215H65ZG
Run Details
Model:
GPT-5.6 Sol (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
51.1s
Cost:
$0.113550
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
99.90%
Dialogue to Total Word Ratio
(details)
99.9507%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 30% 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 blueprints across Bill’s ironing board, because his kitchen table was buried beneath commemorative biscuit tins. The plan showed Saint Barnaby’s community hall, Saturday’s pirate party, and a suspiciously cheerful drawing of balloons. “We enter through the kitchen, blend in, collect presents, and vanish before anyone notices us.” Bill studied the blueprint upside down, nodding with the solemnity of a surgeon gravely examining a sandwich. His disguise lay nearby: striped shirt, eyepatch, hook, and trousers broad enough to conceal three enormous teddy bears. “Small flaw: nobody invited us, you hate children, and my hook squeaks when I walk.” Molly tested the hook. It emitted a rubbery honk, startling a moth from the lampshade. She frowned at it, then drew a moustache on the blueprint’s security guard, who was actually the elderly janitor. “We are not stealing from children. We are temporarily relocating joy into our financial portfolio.” Bill practiced walking. Each step honked, and his foam sword slapped the radiator like an applauding seal. On the third turn, his eyepatch rotated, leaving him blind in both ambition and eyesight for several seconds. “What if they ask which child is ours? We need a believable fictional offspring immediately.” Molly considered this while packing lockpicks inside a cupcake box. Their previous fictional offspring, Algernon, had accidentally aged from six to forty-two during a museum job, largely because Bill had forgotten his own lie again. “Her name is Poppy. She is eight, quiet, gifted, and tragically allergic to follow-up questions.” At noon they approached Saint Barnaby’s dressed as pirates. Molly carried an empty laundry sack labeled TREASURE, while Bill dragged a cardboard ship that caught every doorway, lamppost, bicycle, and one mildly offended nearby terrier. “Walk confidently. Parents never question adults carrying sacks if those sacks have cheerful labels attached.” The hall doors opened before Molly touched them. A gust of shrieking pirates erupted, surrounding Bill’s ship. Tiny hands seized the mast, the sail, and one trouser leg, towing him bodily toward the nearby cake. “Perfect camouflage, Captain Bill. Smile, wave, and stop looking like a tax inspector near treasure.” Bill smiled. Three children screamed. Molly slipped toward the present table, where towers of glittering parcels waited beneath a banner reading HAPPY BIRTHDAY, MAISIE. Her sack opened with the soft, greedy sigh of approaching destiny. “Abort now. Poppy is apparently Maisie’s imaginary enemy, and six witnesses remember her vividly somehow.” Before Molly could answer, a parcel wriggled. She dropped it. The lid sprang off, revealing a white rabbit wearing a pirate hat. It bit her glove, escaped, and charged between Bill’s ankles with criminal enthusiasm. “New plan: rescue rabbit, return presents, accept cake, and never discuss this humiliating afternoon again.” They fled after the rabbit, pursued by delighted children who believed the chase was entertainment. Ten minutes later, Molly sat under the buffet table, icing in her hair, while Bill gently cradled the furry culprit. “I think crime dislikes us.” “Crime has standards. Pass me another sausage roll before surrendering.”
created_at
"2026-07-09T18:05:10.185Z"
7/9/2026, 6:05:10 PM
model
"openai/gpt-5.6-sol-20260709"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3703
moderation_latency
(null)
generation_time
50854
tokens_prompt
77
tokens_completion
1132
native_tokens_prompt
66
native_tokens_completion
3774
native_tokens_completion_images
0
native_tokens_reasoning
3106
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
0.11355
router
(null)
provider_responses
0
endpoint_id
"a54c5de0-89bf-4ad7-a212-cf977eed918a"
id
"resp_0a2f7ea98f22952d016a4fe2d680a08197b6033fe184b16070"
is_byok
false
latency
246
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-1783620310-ozntvsJvfUk9GICLuzKc"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1783620310-LNEbcIIsriC661Ut2nkf"
upstream_id
"resp_0a2f7ea98f22952d016a4fe2d680a08197b6033fe184b16070"
total_cost
0.11355
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
499
99.90%
Dialogue to Total Word Ratio
Ratio: 30.14%, Deviation: 0.14%
neededClean
false
wordsTotal
501
wordsDialogue
151
99.9507%