NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1KA37EN59F7GR5YHXN8B0V6
Run Details
Model:
DeepSeek V4 Pro 0813 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
2m 45s
Cost:
$0.043429
Provider:
StreamLake
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
95.89%
Dialogue to Total Word Ratio
(details)
97.9438%
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
Molly squinted through the grimy windshield at the inflatable dragon guarding the community center entrance. The dragon listed slightly to the left, its nylon claws scuffing the pavement. She tapped a glittery invitation against her knee, the corner already soft from her thumb. "Okay, run it again. Bounce house at two." Bill, wedged into the driver's seat with a party hat strapped over his beanie, frowned at a napkin diagram of the venue. The van smelled of stale fries and the cinnamon air freshener shaped like a cartoon fox. A red crayon X marked the gift table. A blue squiggle represented the llama pen. He rotated the napkin twice before speaking. "Why can't we just walk in as clowns? Clowns are invisible at kids' parties." Molly grabbed the napkin and smoothed it against her thigh. "You're six-five with a neck tattoo. We go in as parents. Sign the sheet, say Liam's folks send love." She traced a route from the entrance to the corner where the presents would be stacked like a glittering, ungrateful tower. Her finger lingered on the fire alarm. Bill nodded, then paused. "Which one is Liam?" "Birthday boy. His mom hired a llama, so the pile will be huge." Molly said it without looking up. She was already picturing the diaper bag, the extra-large gift sack with reinforced handles, the ninety-second exit. Outside, a mother dragged a toddler toward the entrance, the child clutching a present almost as large as himself. Bill tapped the steering wheel. "What if the llama recognizes me?" Molly stared at the inflatable dragon. It wobbled in the January wind, its painted smile somehow knowing. "It's not a security llama, Bill. It's a petting zoo rental." She folded the napkin into a small, sharp rectangle. He considered this, then examined the side door as if it might bite. "What if I get stuck in a bounce house?" "You won't hide in a bounce house. Stand by the door and look panicked." She reached into the back seat and pulled out a lanyard reading VOLUNTEER. The letters were slightly askew. The lanyard had been borrowed from a blood drive three months earlier. Bill adjusted his party hat. "So we clear about thirty gifts? That's maybe four hundred bucks on resale." "Easily. There's a tablet shaped like a crayon box. Saw it on the registry." She looped the lanyard over her neck. The plastic badge bounced against her sternum. Her jacket already held two empty pillowcases and a roll of duct tape. Bill looked at the dragon again. The party hat slipped over one eye. He pushed it back with his thumb. His fingers drummed on the steering wheel. "One more thing. What's the code word if things go bad?" Molly opened the door. Cold air rushed in, carrying the distant thump of a party DJ. She stepped onto the curb, then turned back. The napkin slipped from her pocket and skittered under the van. "Use your grown-up words, Bill. Just say 'llama.'"
created_at
"2026-09-03T09:37:42.882Z"
9/3/2026, 9:37:42 AM
model
"deepseek/deepseek-v4-pro-20260813"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1582
moderation_latency
(null)
generation_time
164745
tokens_prompt
75
tokens_completion
11938
native_tokens_prompt
146
native_tokens_completion
12930
native_tokens_completion_images
(null)
native_tokens_reasoning
12265
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
0.043429214
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788428262-hYoNlxprnYw6UYupvikt"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788428262-bjurhHZteK3DYTZH2IqW"
upstream_id
"req-5hrv82-1788428263049754256"
provider_responses
0
endpoint_id
"072c1c23-c069-4242-bb6d-1b30ca5f4eb8"
id
"req-5hrv82-1788428263049754256"
is_byok
false
latency
1581
model_permaslug
"deepseek/deepseek-v4-pro-20260813"
provider_name
"StreamLake"
status
200
total_cost
0.043429214
cache_discount
(null)
upstream_inference_cost
0
provider_name
"StreamLake"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
100.00%
Matches word count
n/a
neededClean
false
words
499
95.89%
Dialogue to Total Word Ratio
Ratio: 29.08%, Deviation: 0.92%
neededClean
false
wordsTotal
502
wordsDialogue
146
97.9438%