NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-16-01KHM5JAKNS77YJFC775B5FMA9
Run Details
Model:
DeepSeek V3 (2025-03-24)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
29.6s
Cost:
$0.000466
Provider:
ModelRun
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 70% 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
**Scene: The Hideout** The dim glow of a flickering bulb cast long shadows across the cluttered hideout—a musty basement with stacks of old newspapers, a dented coffee pot, and one very nervous-looking man in a clown costume. Molly tapped her fingers against a hand-drawn blueprint labeled *"BIRTHDAY BONANZA"* in messy cursive. "Alright, Bill, let’s run it one more又何时间." Bill adjusted hisغيرة红鼻子 and groaned. "We’ve run it twelve times, Molly. It’s a kid’s party, not Fort Knox." Molly shot him a look. "Exactly why we can’t mess this up. If we’re wrong, we get taken down by an army of six-year-olds hopped up on cake." Bill rolled his eyes. "Oh no, they’ll hit us with their foam swords and sticky fingers!" Molly pointed firmly at the blueprint. "First, you go in dressed as Clumsy Clive, the party clown—" "—the worst alias ever—" "—distract the kids with your ‘magic’ balloon animals—" "Which I can’t make!" "Just tie knots and call it modern art. Meanwhile, I’ll sneak into the backyard where they stash the presents." Bill frowned. "What if someone notices you?" Molly smirked. "I’ll be disguised as the overly enthusiastic aunt. No one questions auntie—she pinches cheeks and asks about homework. Everyone avoids her." Bill sighed. "And the getaway?" "Same as always. We meet at the inflatable slide, dive in, and roll out the back fence." Bill blinked. "That’s our genius plan? Clown and auntie dive into a slide?" Molly shrugged. "Sometimes simplicity’s gold, Bill. Plus, we’ve got the perfect cover—if anything goes wrong, we just yell ‘surprise!’ and start a conga line. No one suspects the conga." Bill paused, then slowly nodded. "Alright. But if we end up in a piñata, I’m blaming you." Molly grinned. "Deal. Now go practice your clown walk." Bill took two wobbly steps before tripping over his oversized shoes. Molly sighed. "Or just oxford stumble dramatically. That works too." Bill groaned from the floor. "We’re gonna die by glitter." Molly offered a hand to pull him up. "But we’ll die *rich* in finger paint and LEGOs." Bill considered it, then grinned. "Alright, let’s go rob a birthday." They bumped fists. The birthday cake didn’t stand a chance.
created_at
"2026-02-16T21:27:10.136189+00:00"
2/16/2026, 9:27:10 PM
model
"deepseek/deepseek-chat-v3-0324"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2243
moderation_latency
(null)
generation_time
29607
tokens_prompt
77
tokens_completion
627
native_tokens_prompt
66
native_tokens_completion
588
native_tokens_completion_images
(null)
native_tokens_reasoning
0
native_tokens_cached
5
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
0.00046596
router
(null)
provider_responses
0
id
"chatcmpl-7552cdac8a5540a3a11ebbd840636343"
status
200
is_byok
false
latency
90
endpoint_id
"a3938710-1b99-4a1f-85f4-2494f3f659d3"
provider_name
"ModelRun"
model_permaslug
"deepseek/deepseek-chat-v3-0324"
api_type
"completions"
id
"gen-1771277200-ynl8m4UEjBjlcAhwbNFM"
upstream_id
"chatcmpl-7552cdac8a5540a3a11ebbd840636343"
total_cost
0.00046596
cache_discount
(null)
upstream_inference_cost
0
provider_name
"ModelRun"
Evaluation details
Result
Evaluator
Details
Meta Data
0.0000%
Matches word count
n/a
neededClean
false
words
370
4.6348%
Dialogue to Total Word Ratio
Ratio: 62.16%, Deviation: 7.84%
neededClean
false
wordsTotal
370
wordsDialogue
230
2.3174%