NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-04-25-01KQ2H2VB7PCB3N5CVF9VZ01KJ
Run Details
Model:
DeepSeek V4 Pro
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
28.8s
Cost:
$0.000843
Provider:
DeepSeek
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
0.0000%
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
The warehouse smelled of forgotten balloon animals and industrial-strength cleaner, a scent Molly found aggressively nostalgic. Across the folding table, Bill was trying to unpick a complicated padlock with a pair of tweezers, a popsicle stick, and a worrying amount of optimism. Practice, he’d called it. “You know, a normal person would just bring a gift,” Molly said, not looking up from the architectural blueprint of a bouncy castle she was annotating. Bill squinted, his tongue peeking from the corner of his mouth. “A normal person pays retail for a doll that wets itself on a timer. We are artists. Also, that lock on the gymnasium side-door is a joke. I could open it with a stern look.” “I’ve seen you try to open a bag of chips with a stern look. It ended in tragedy and a trip to the emergency room.” She drew a small, aggressive ‘X’ on a kidney-shaped bounce chamber. “My point is, the presents are the easy part. The containment system is the problem.” “Containment system? Molls, it’s a bouncy castle. Full of kids. Hyped-up, cake-fueled, pre-crash anarchy machines.” “Exactly,” she said, tapping the blueprint. “This is Fort Knox made of pressurized vinyl. Timmy the Terrific is having his seventh birthday. His mother, an amateur chemist of passive aggression, hired a clown. A full-face-paint, rainbow-suspender, silent-terror clown. Our window is during the clown show. Thirty minutes while the sugar-gremlins are hypnotised by a man making a giraffe out of a single balloon. The parents will be mainlining cheap chardonnay, grateful for the break.” Bill finally snapped the padlock open with a satisfying click and beamed like he’d just solved cold fusion. “So I walk in, grab the heap of expensive, brightly-colored future landfill, and walk out. A ghost amongst the juice boxes.” “A ghost wearing a tactical vest lined with garbage bags? No. You’ll be dressed as the party’s second-tier entertainment. Sparkles the Unicorn.” Bill’s smile evaporated. “Absolutely not. The last time I wore a costume, I still have a rash in a place that suggests a deep betrayal by synthetic fur.” “It’s a non-negotiable costume requirement. You’re a decoy. You wade into the fray, a majestic, horned fool, and create a distraction. Any kid still teetering on the edge of clown-hypnosis gets pulled into your orbit.” “A distraction for what?” Molly pulled a device from her bag that looked like a fishing rod mated with a vacuum cleaner attachment. “While you’re doing spirit fingers and ‘whimsical’ neighing, I’ll be in the ceiling. I’ve located a crawl space above the gym. I pop a ceiling tile, and this,” she said, patting the device, “is a custom-built, silent, child-proof gift extraction arm. I lower the claw, snag a present, retract. A quiet, vertical conveyor belt of greed. The kids think the presents are magically floating up to Unicorn Heaven.” Bill stared at her, a muscle twitching in his jaw. “The ceiling. You get to be the phantom creature of stealthy logic, playing a surgical claw-game, and I have to be a mylar-hoofed cartoon on the ground, getting pelted with half-eaten fruit snacks?” “You’re better at jazz hands,” she said flatly. “And you glow under blacklight. It’s your gift. Now, the critical path is the extraction. We can’t just leg it across the soccer field with fifty loot bags. I’ve commandeered the school’s old laundry cart. The one on the squeaky wheel that nobody uses because it sounds like a dying goose. I’ve greased it, and it’s our getaway vehicle. I’ll zip-line the haul down from the roof access door.” “A laundry cart? Why not just get an actual van?” “Because, my sweet Bill, a van is suspicious. A laundry cart pushed by a distraught unicorn muttering about a broken dream is merely a tragedy. People will look away out of pity. It’s the perfect, pathetic escape. We blend into the mundane sadness of a party entertainer’s failed life.”
created_at
"2026-04-25T14:35:46.923Z"
4/25/2026, 2:35:46 PM
model
"deepseek/deepseek-v4-pro-20260423"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
853
moderation_latency
(null)
generation_time
28765
tokens_prompt
77
tokens_completion
1047
native_tokens_prompt
67
native_tokens_completion
935
native_tokens_completion_images
(null)
native_tokens_reasoning
0
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"
usage
0.000842595
router
(null)
provider_responses
0
endpoint_id
"d084170b-dd99-4e53-9ae1-b63881f3cfbf"
id
"5e67d98e-96f0-4963-86c3-69679d1c748a"
is_byok
false
latency
280
model_permaslug
"deepseek/deepseek-v4-pro-20260423"
provider_name
"DeepSeek"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.1; linux; x64))"
http_referer
(null)
request_id
"req-1777127746-xoidVihE14V07hFBqKNC"
session_id
(null)
api_type
"completions"
id
"gen-1777127746-KtpKPRLaAgq3RqiYzbJq"
upstream_id
"5e67d98e-96f0-4963-86c3-69679d1c748a"
total_cost
0.000842595
cache_discount
(null)
upstream_inference_cost
0
provider_name
"DeepSeek"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
652
0.00%
Dialogue to Total Word Ratio
Ratio: 78.87%, Deviation: 48.87%
neededClean
false
wordsTotal
672
wordsDialogue
530
0.0000%