NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1K9PJW99GNHDSAFXK3DE23Q
Run Details
Model:
DeepSeek V4 Flash 0731 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 200 words with 10% dialogue
Duration:
5.5s
Cost:
$0.000578
Provider:
AtlasCloud
Evaluation summary
Score
Evaluator
Details Link
14.30%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
7.1507%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 200 word scene that has 10% 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 flattened the party invitation against the dashboard, her finger tracing the grinning clown. “We hit the bounce castle at three.” Bill squinted at the suburban house across the street. A piñata hung from the maple like a fleshy piñata, and a mountain of wrapped gifts clustered under a banner reading “HAPPY 7TH BIRTHDAY, KEVIN.” He adjusted his fake mustache. “The parents look like the type who buy gift receipts.” “Gift receipts are for amateurs,” Molly said. “We need the whole haul. The Lego set alone is worth four hundred on the dark web.” Bill checked his watch, then the sky, then the sandwich he’d packed. “What’s the plan for the kids?” “Cry. Point at the sky. Say a bird stole their dad’s car.” Molly shrugged. “They’re seven. They believe everything.” “And the parents?” “You cause a diversion. I’ll do the lifting.” “What kind of diversion?” Molly grinned. “You tell them you’re a magician and ask for a volunteer. Then you make their toddler disappear.” Bill nodded slowly. “Into the bounce castle.” “Into the bounce castle.” The front door opened. A woman in a party hat stepped out, yelling at a man about the ice cream melting. Molly popped the car door and slipped a ski mask over her face. “Showtime, partner.” Bill sighed. “I never got to eat my sandwich.”
created_at
"2026-09-03T09:30:48.613Z"
9/3/2026, 9:30:48 AM
model
"deepseek/deepseek-v4-flash-20260731"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
857
moderation_latency
(null)
generation_time
5356
tokens_prompt
75
tokens_completion
435
native_tokens_prompt
146
native_tokens_completion
389
native_tokens_completion_images
(null)
native_tokens_reasoning
74
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
0.00057772
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788427848-OyhHwa2HkkOmhrGJOV9N"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788427848-zJFAQxD56rcYqHgFyb8e"
upstream_id
"34801da4aa2d45778f2783c11aeb091f"
provider_responses
0
endpoint_id
"35b202d3-9dbe-4367-9bc1-f2b327472493"
id
"34801da4aa2d45778f2783c11aeb091f"
is_byok
false
latency
857
model_permaslug
"deepseek/deepseek-v4-flash-20260731"
provider_name
"AtlasCloud"
status
200
total_cost
0.00057772
cache_discount
(null)
upstream_inference_cost
0
provider_name
"AtlasCloud"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
14.30%
Matches word count
n/a
neededClean
false
words
221
0.00%
Dialogue to Total Word Ratio
Ratio: 52.04%, Deviation: 42.04%
neededClean
false
wordsTotal
221
wordsDialogue
115
7.1507%