NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-04-25-01KQ2EPRYSQ03GCHE92T2D06ZH
Run Details
Model:
DeepSeek V4 Flash (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
5m 37s
Cost:
$0.000253
Provider:
DeepInfra
Evaluation summary
Score
Evaluator
Details Link
68.10%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
34.0511%
Overall
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
“You sure about this, Bill?” Molly asked, peering through the binoculars at the bounce house. “A seven-year-old’s birthday party? We’ve knocked over banks.” “Banks have alarms, guards, and bulletproof glass,” Bill said, adjusting his ridiculous sunglasses. “This has a piñata and a lady in a unicorn onesie. Low risk, high reward. Look at those presents. That’s not just LEGOs and slime. That’s a Nintendo Switch box. Maybe two.” “And the parents? They’re just going to let us waltz in with a crowbar?” “We’re not using a crowbar. We’re using the ‘official party entertainer’ cover. I’ve got the balloon animals. You’ve got the makeup. We walk in, say we’re from ‘Clowning Around Inc.,’ do two tricks, and while they’re focused on the disappearing handkerchief, you’ll be stuffing gift bags into the laundry cart.” Molly lowered the binoculars. “The laundry cart? We’re pushing a laundry cart through a backyard full of toddlers?” “It’s on wheels. It’s beige. It’s invisible.” “It’s a laundry cart, Bill. With a giant ‘SUDSY’ logo on the side. People will notice.” “They’ll think we’re the cleaning crew. Parents love cleaning crews. It means the cake crumbs are handled.” She sighed. “Okay, let’s walk through it. I’m the clown. You’re the assistant. I do the rabbit trick. You’re holding the hat. Where’s the rabbit, by the way?” “We don’t have a rabbit. We have a gerbil.” “A gerbil? You want to pull a gerbil out of a hat at a kid’s party?” “It’s fluffy. They won’t know. And if it bites me, I’ll scream and they’ll feel sorry for me. Distraction.” “What about the parents who hired the real entertainer? The kids are going to ask about ‘Mr. Wiggles’.” “Mr. Wiggles is a myth. I looked him up. He’s booked solid. They’ll be happy to see anyone. Besides, I have a fake mustache. I look trustworthy.” “You look like a janitor who got lost on his way to a 1980s cop movie.” “Thank you. That’s the vibe. Trustworthy but firm. Now, the exit strategy: we load the cart, I drop a smoke bomb near the grill, everyone panics about the hamburgers, we exit through the side gate.” “And if the birthday girl starts crying?” “We give her a balloon. A sad one. She’ll forget.” “She won’t forget. She’ll remember the weird clown and the guy with the gerbil who took her gifts.” “Perfect. She’ll have a story for therapy. Look, Molly, we’re not villains. We’re redistributing consumer goods. Those kids have too much plastic crap. We’re doing a public service.” “We’re stealing from children.” “We’re stealing from the parents who bought too much. The kids won’t even notice until bedtime. By then, we’re long gone. Come on. Just one job. Then we’re out of the game.” Molly rubbed her temples. “Fine. But if that gerbil bites me, I’m using the smoke bomb on your face.” “Deal. Now let’s go be the life of the party.”
created_at
"2026-04-25T13:54:14.11Z"
4/25/2026, 1:54:14 PM
model
"deepseek/deepseek-v4-flash-20260423"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
530
moderation_latency
(null)
generation_time
337118
tokens_prompt
77
tokens_completion
927
native_tokens_prompt
67
native_tokens_completion
870
native_tokens_completion_images
(null)
native_tokens_reasoning
104
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
0.00025298
router
(null)
provider_responses
0
endpoint_id
"934a69f9-bd54-474b-beca-24560f721e12"
id
"1f331de5eeab41c2b5b5b04b5be8fb53"
is_byok
false
latency
53
model_permaslug
"deepseek/deepseek-v4-flash-20260423"
provider_name
"DeepInfra"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.1; linux; x64))"
http_referer
(null)
request_id
"req-1777125254-jAF4CtlULtOdF0GJ2oy6"
session_id
(null)
api_type
"completions"
id
"gen-1777125254-KokfKE9OngWZUlyOC4vC"
upstream_id
"1f331de5eeab41c2b5b5b04b5be8fb53"
total_cost
0.00025298
cache_discount
(null)
upstream_inference_cost
0
provider_name
"DeepInfra"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
68.10%
Matches word count
n/a
neededClean
false
words
486
0.00%
Dialogue to Total Word Ratio
Ratio: 94.67%, Deviation: 24.67%
neededClean
false
wordsTotal
488
wordsDialogue
462
34.0511%