NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-16-01KHM5DYGYD8WK99KH2K7FCWR0
Run Details
Model:
DeepSeek V3 (2024-12-26)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
24.6s
Cost:
$0.000694
Provider:
DeepInfra
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
The sun dipped low in the sky, casting a golden hue over the abandoned warehouse where Molly and Bill sat hunched over a map of the neighborhood. A single flickering bulb swung above them, casting erratic shadows on the walls. Molly tapped the map with her pen, her fiery red hair bouncing as she leaned forward. “Alright, Bill, here’s the plan,” she said, her voice low but brimming with excitement. “The party’s at the Anderson place. Starts at noon. They’re gonna have a bouncy castle, a clown, and—get this—a piñata shaped like a unicorn. That thing’s probably stuffed with candy, but we’re not here for the candy.” Bill squinted at the map, scratching his scruffy beard. “So we’re robbing a kid’s birthday party? Molly, I thought we were professionals. This is... kinda low, even for us.” Molly rolled her eyes and flicked his arm. “Oh, come on, Bill. Think about it. Parents drop hundreds of dollars on those gifts. We’re talking smartphones, gaming consoles, designer toys—all wrapped up in shiny paper, begging to be liberated. And the best part? Everyone’s distracted by the clown. No one’s gonna notice us slipping in and out.” Bill leaned back in his creaky chair, folding his arms. “And how, exactly, do you plan on sneaking into a backyard full of screaming kids and hyperactive dogs?” Molly grinned, pulling out two costumes from a bag behind her. One was a rainbow-colored clown suit complete with oversized shoes and a curly wig. The other was a unicorn onesie with a glittery horn. “We’re going undercover,” she said, tossing the unicorn outfit at Bill. “You’re the unicorn. I’m the clown. We blend in, grab the gifts, and disappear before anyone even notices.” Bill stared at the unicorn costume, his expression somewhere between disbelief and horror. “You’ve got to be kidding me. I’m not wearing that.” Molly wagged a finger at him. “Oh, yes, you are. Unless you’ve got a better idea?” Bill sighed, pinching the bridge of his nose. “Fine. But if anyone takes a picture of me in that thing, I’m outing you to the cops.” Molly laughed, clapping her hands. “Deal! Now, here’s the timeline: we arrive at noon, sneak in during the piñata chaos, grab the gifts from the designated ‘present table,’ and stash them in the clown car—which, by the way, I’ve already procured.” Bill raised an eyebrow. “Clown car? You didn’t think to mention that earlier?” “Details, details,” Molly said, waving him off. “The car’s parked two blocks away. We load up, ditch the costumes, and split the loot. Easy peasy.” Bill muttered under his breath, “Easy peasy. Right. Just your average Tuesday.” Molly stood, draping the clown suit over her arm. “Cheer up, Bill. By this time tomorrow, we’ll be sitting on a mountain of presents. And hey, if we’re lucky, maybe the piñata will have cash in it.” Bill snorted, reluctantly pulling on the unicorn costume. “You’ve got a twisted sense of ambition, you know that?” Molly grinned, adjusting her curly wig. “That’s why you love me, Bill. Now let’s go make some kids cry.”
created_at
"2026-02-16T21:24:41.553629+00:00"
2/16/2026, 9:24:41 PM
model
"deepseek/deepseek-chat-v3"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
334
moderation_latency
(null)
generation_time
24612
tokens_prompt
77
tokens_completion
833
native_tokens_prompt
66
native_tokens_completion
756
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)
origin
"https://ncbench.com/"
is_byok
false
finish_reason
"stop"
native_finish_reason
"stop"
usage
0.00069396
router
(null)
provider_responses
0
id
"chatcmpl-Rer9qQD3k9G41lqvrORzrQZI"
status
200
is_byok
false
latency
50
endpoint_id
"5294d55f-9012-496b-8f22-8cc919432dcd"
provider_name
"DeepInfra"
model_permaslug
"deepseek/deepseek-chat-v3"
api_type
"completions"
id
"gen-1771277056-GIYy05CZtnUJ65re58r7"
upstream_id
"chatcmpl-Rer9qQD3k9G41lqvrORzrQZI"
total_cost
0.00069396
cache_discount
(null)
upstream_inference_cost
0
provider_name
"DeepInfra"
Evaluation details
Result
Evaluator
Details
Meta Data
20.1897%
Matches word count
n/a
neededClean
false
words
520
0.3722%
Dialogue to Total Word Ratio
Ratio: 60.58%, Deviation: 10.58%
neededClean
false
wordsTotal
520
wordsDialogue
315
10.2809%