NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-09-01KX40AHNDK55SWKZR27ZQ39H1
Run Details
Model:
GPT-5.6 Sol (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
44.8s
Cost:
$0.122520
Provider:
OpenAI
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
100.00%
Dialogue to Total Word Ratio
(details)
100.0000%
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
Molly spread the stolen floor plan across a cafe table, pinning each corner with sugar bowls while Bill nervously buttered a napkin at her side. “The birthday girl turns six at three o’clock. By three fifteen, every adult will be distracted by cake, candles, and competitive photography near the balloons.” Bill raised the napkin, discovered what he had done, and quietly folded the greasy evidence into a swan with deeply compromised structural integrity and dignity. “You’re forgetting the clown. Clowns notice everything. They’re painted like fools specifically so nobody suspects their advanced powers of observation and balloon surveillance remarkably well.” Molly tapped a red circle marked ENTERTAINER, then slid Bill a photograph of a cheerful man wearing trousers broad enough to conceal Luxembourg or Belgium. “That is Mr. Giggles. He arrives at two forty. You will intercept him outside, borrow his costume, and enter through the kitchen door carrying balloons.” Bill stared at the photograph. Mr. Giggles had orange eyebrows, green shoes, and the haunted expression of someone who had swallowed a bicycle horn sideways. “I cannot impersonate a clown. I have standards, arches, and a medically documented fear of face paint. Also, those shoes clash with my soul violently.” Outside, a bus sighed at the curb. Inside, Molly calmly drew a blue arrow from the kitchen toward a room labeled PRESENT FORTRESS in crayon. “The mother stores gifts in the study until games finish. We cross the hall, fill two sacks, then escape through the laundry chute before supper.” Bill peered closer. The laundry chute ended in a basement beside a furnace, three recycling bins, and something Molly had illustrated as fangs with whiskers. “What is that?” “The family dog. His name is Professor Nibbles, and according to my source, he dislikes clowns and burglars, especially near snack time.” The napkin swan slowly unfolded itself, as if unwilling to remain at the meeting. Bill watched it collapse and seemed jealous of its freedom entirely. “Your source is seven, wears a pirate hat, and traded you this intelligence for jellybeans. Perhaps we should question the reliability of our informant slightly.” Molly produced another photograph. It showed a solemn child with an eyepatch, holding a juice box like a witness under government protection at recess today. “Captain Archie is trustworthy. He also says the presents include a telescope, a robot, and three hundred pounds in untraceable building blocks, plus stickers too.” Bill stopped buttering things. For the first time, greed brightened his face, although caution remained perched behind it, wearing a tiny reflective vest and helmet. “Building blocks are difficult to fence. Every buyer asks questions: Are they complete? Are they clean? Why does this astronaut smell so extremely funny, anyway?” Molly leaned forward, lowering her voice as a waitress passed. The waitress glanced at the plans, the photographs, and Bill’s buttery origami debris with concern. “We are not fencing anything. We sell everything back to the parents after the children unwrap it. Convenience fee. Environmentally friendly. Practically heroic and educational.”
created_at
"2026-07-09T17:55:33.172Z"
7/9/2026, 5:55:33 PM
model
"openai/gpt-5.6-sol-20260709"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
35927
moderation_latency
(null)
generation_time
44622
tokens_prompt
77
tokens_completion
1101
native_tokens_prompt
66
native_tokens_completion
4073
native_tokens_completion_images
0
native_tokens_reasoning
3435
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
0.12252
router
(null)
provider_responses
0
endpoint_id
"a54c5de0-89bf-4ad7-a212-cf977eed918a"
id
"resp_08572aefdbe06234016a4fe0953cf8819084ed780cbfa43a0f"
is_byok
false
latency
203
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-1783619733-Xk5KvEqLpnB1Vj4MVezI"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1783619733-degu7WOgMPXAlMcoAygq"
upstream_id
"resp_08572aefdbe06234016a4fe0953cf8819084ed780cbfa43a0f"
total_cost
0.12252
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: 50.00%, Deviation: 0.00%
neededClean
false
wordsTotal
500
wordsDialogue
250
100.0000%