NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-10-01M25DV2XQ0087CTMH1XRDRXMG
Run Details
Model:
DeepSeek V4.1 Flash (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
44.8s
Cost:
$0.006639
Provider:
DeepSeek
Evaluation summary
Score
Evaluator
Details Link
99.38%
Matches word count
(details)
99.95%
Dialogue to Total Word Ratio
(details)
99.6645%
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 birthday invitation across the motel bed, its cartoon dinosaur grinning up at them like it knew something. She uncapped a red marker and drew a circle around the address. Bill sat opposite, eating a gas-station sandwich and looking like a man who had already been arrested twice in his imagination. Rain tapped the window. Somewhere down the hall, a vacuum cleaner howled. "The Wigglesworth twins turn seven Saturday," Molly said. "Forty children, two clowns, and a bounce house shaped like a castle." Bill swallowed. "And how many presents?" "Thirty-seven," Molly said. "All in the conservatory, wrapped, labeled, and guarded by a nine-year-old named Timothy." "Timothy," Bill repeated. "Is he armed?" "He has a whistle and a clipboard." "We're doomed." Molly tapped the invitation. "We're efficient. Timothy gets the remote-control dinosaur from the top of the pile. He becomes our lookout." Bill rubbed his temples. "You want to bribe a child with stolen merchandise?" "It's not stolen yet. That's what makes it ethical." He stared at her. She ignored him and unfolded a diagram of the party venue, complete with arrows, timings, and a sad drawing of a clown. "Cake is at three fifteen. Parents will be filming. Kids will be screaming. You will be doing bubbles." "Why am I always the bubble man?" "Because you fit inside the costume." "I'm six feet tall." "And the costume is forgiving." Bill sighed. "What's my exit?" "The magician's van. My cousin Derek owes me." "Does Derek know he's a getaway driver?" "He thinks he's delivering balloons." Molly folded the diagram and slid it into her sleeve. Bill finished his sandwich, crumpled the wrapper, and missed the trash can by a foot. "Fine," Bill said. "But if a clown asks why I'm stuffing teddy bears into a laundry cart, I'm blaming you." Molly smiled. "Tell him it's a magic trick. By the time he understands, we'll be gone." Bill picked up his coat. "One more question." "What?" "Why children's parties?" "Because adults have security cameras," Molly said. "Children only have frosting." Bill considered this. "And the parents?" "Ball pit. Cake coma. Group photos." "If anyone asks, we're from the party planning committee." "With lanyards and everything." "Lanyards make people trust you." "That's how I got into the aquarium." "I thought that was the penguin." "The penguin was a complication." "We'll be legends." "Or arrested." "Both." "And rich in teddy bears." Molly pulled two laminated badges from her pocket. They said PARTY LOGISTICS in cheerful, threatening letters. Bill clipped his to his collar and tried to look official, which mostly made him look like a tax auditor at a bouncy castle. They shook hands. Then Bill remembered the aquarium, the penguin, and the restraining order, and he shook his head instead. Molly drew a tiny party hat on the diagram, because even criminal masterminds appreciate branding. Outside, an ice-cream truck played a tune that sounded suspiciously like a warning. Bill put on a fake mustache. Molly put on a fake smile. They practiced saying delightful in unison.
created_at
"2026-09-10T10:29:30.172Z"
9/10/2026, 10:29:30 AM
model
"deepseek/deepseek-v4.1-flash-20260910"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
715
moderation_latency
(null)
generation_time
44759
tokens_prompt
75
tokens_completion
10297
native_tokens_prompt
93
native_tokens_completion
11041
native_tokens_completion_images
(null)
native_tokens_reasoning
10343
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
"stop"
service_tier
(null)
usage
0.00663855
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.20.0; linux; x64))"
http_referer
(null)
request_id
"req-1789036170-3nnVgu8uYYlwT1E6lA9l"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1789036170-Q6Bnh8tAHphP7Huqi7eR"
upstream_id
"0d176cb5-29e3-4425-97a9-eac97bf7eba2"
provider_responses
0
endpoint_id
"fd104bc8-535a-4d31-a847-a8e402e86a02"
id
"0d176cb5-29e3-4425-97a9-eac97bf7eba2"
is_byok
false
latency
329
model_permaslug
"deepseek/deepseek-v4.1-flash-20260910"
provider_name
"DeepSeek"
status
200
total_cost
0.00663855
cache_discount
(null)
upstream_inference_cost
0
provider_name
"DeepSeek"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
99.38%
Matches word count
n/a
neededClean
false
words
505
99.95%
Dialogue to Total Word Ratio
Ratio: 49.90%, Deviation: 0.10%
neededClean
false
wordsTotal
511
wordsDialogue
255
99.6645%