NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1KCM1HE5QA2MC1QT1DED5JY
Run Details
Model:
DeepSeek V4 Pro 0813 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
12.4s
Cost:
$0.003333
Provider:
Parasail
Evaluation summary
Score
Evaluator
Details Link
0.01%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
0.0049%
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 stared at the finger-painted banner sagging across the community center doors. “Happy 6th Birthday, Timmy!” it declared, above a drawing of a dinosaur that looked vaguely ashamed of itself. “I can’t believe we’re doing this,” Bill whispered, adjusting his oversized sunglasses. “We’ve pulled job at a casino, a private zoo, and once, memorably, a cheese vault. But this? This is a new low.” “It’s not a low, Bill. It’s a pivot.” Molly tugged at the collar of her rented clown costume. The ruffled neck itched like a nest of angry bees. “Children’s parties have one thing casinos don’t. A mountain of unguarded merchandise wrapped in shiny paper.” “And a bouncy castle full of human missiles,” Bill muttered. “What’s the play? Distract them with balloon animals while you stuff a sack into a stroller?” “Better.” Molly pointed toward the side entrance, where a teenager in a cheap pirate hat was sneaking a vape. “That’s the entertainment. Cap’n Chuckles. He’s supposed to lead a treasure hunt in ten minutes. All the kids go out back. Presents sit alone, vulnerable, on the gift table.” Bill squinted. “And we are?” “You’re the backup magician. I’m the face painter.” She handed him a sparkly cape. “We slip in. You pull a quarter from the birthday boy’s ear. I paint a unicorn on a toddler. Then, when the treasure hunt starts, we clear the table.” Bill draped the cape over his shoulders. It said “World’s Okayest Uncle” in glitter. “What if a kid sees us taking the gifts?” “We tell them Santa’s doing an emergency inventory audit. Kids trust adults in sad costumes. It’s practically a law.” “And if a parent asks who invited us?” Molly grinned. “We’re from the party planning app. Very common. We brought our own brushes. Very professional.” Bill sighed. “I hate that this might work. But what about the piñata? That’s a candy-filled wildcard. One good swing and we’re showered in sugar and witnesses.” “We’ll hide behind the cake table during the piñata. Classic cover. Buttercream and chaos.” “So the plan is: sparkle, lie, steal, hide behind a sheet cake shaped like a dump truck.” Bill shook his head. “This is either genius or the exact reason we’ll end up in a very small jail cell with a ball pit.” Molly checked her watch. “Genius. Now waddle. Clowns don’t stride. They waddle with purpose.” They shuffled toward the door, capes fluttering, as the first child spotted them. “Look, Mommy!” the child shrieked. “A sad magician and a clown with a wristwatch!” Bill forced a smile. “She’s not sad, kid. She’s method.” “That means she’s very serious about stealing your presents,” Molly whispered. Bill elbowed her. “Not out loud, Molly. We’re inside the perimeter.” “Relax, Bill. Nobody listens to clowns.” She waved at the child. “Especially not the honest ones.”
created_at
"2026-09-03T10:21:51.038Z"
9/3/2026, 10:21:51 AM
model
"deepseek/deepseek-v4-pro-20260813"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
757
moderation_latency
(null)
generation_time
12305
tokens_prompt
75
tokens_completion
867
native_tokens_prompt
146
native_tokens_completion
793
native_tokens_completion_images
(null)
native_tokens_reasoning
93
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
0.003333
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788430911-AndHPd3sB4vttiYzrTWc"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788430911-Uk0PnnNLUBvVEWgQmcsO"
upstream_id
"1300295b3eb24881930e774d9a2cea4c"
provider_responses
0
endpoint_id
"487a0ef9-ddaa-4d3d-b882-7912daa08555"
id
"1300295b3eb24881930e774d9a2cea4c"
is_byok
false
latency
757
model_permaslug
"deepseek/deepseek-v4-pro-20260813"
provider_name
"Parasail"
status
200
total_cost
0.003333
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Parasail"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
0.01%
Matches word count
n/a
neededClean
false
words
469
0.00%
Dialogue to Total Word Ratio
Ratio: 71.55%, Deviation: 21.55%
neededClean
false
wordsTotal
471
wordsDialogue
337
0.0049%