NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1KBAZ51YM863MKMFXZ4N4EB
Run Details
Model:
DeepSeek V4 Pro 0813 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
45.1s
Cost:
$0.012136
Provider:
SiliconFlow
Evaluation summary
Score
Evaluator
Details Link
3.62%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
1.8117%
Overall
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
Molly adjusted the clown nose on the dashboard and sighed. "We are not going in as clowns." Bill looked up from a map drawn on a napkin. "Why not? It's a children's party. Clowns are invisible." "Clowns are terrifying. One kid screams, every parent looks, and we're caught with our hands in the gift pile." She tapped the invitation. "We go as caterers." "Caterers don't wear ski masks." "We won't wear ski masks. We'll wear aprons and those little white hats. I already ironed them." Molly pulled two chef hats from a duffel bag. Bill frowned. "You said this would be dignified." "I said profitable. There's a bouncy castle, a magician, and forty-seven presents. The birthday boy is six. His parents have more money than sense. We slide in, load the gift table into laundry carts, and slide out during the puppet show." "What if a kid asks us for more juice?" Molly handed him a juice box. "Give them this. It's spiked with nothing. They'll love you." Bill took the juice box and held it like a grenade. "And the magician? He'll notice two strangers hauling presents." "That's why we bring a third laundry cart and tell him we're collecting the dirty tablecloths. Magicians hate laundry. It ruins the drama." Molly checked her watch. "We have twenty minutes until the piñata breaks. That's our window." "Why the piñata?" "Because when the candy hits the grass, every adult becomes a referee and every kid becomes a feral raccoon. Chaos covers all sins." She started the van. "Put on the hat." Bill put on the chef hat. It slumped to one side. "I feel like a muffin." "You look like a man who belongs in a kitchen. Now, the code word is 'allergic to peanuts.' If I say it, we abort. If you say it, I'll pretend I didn't hear you." "That's a terrible system." "It's a heist, Bill. Not a marriage." Molly pulled out of the parking lot. "When we get there, you take the left side of the gift table. I'll take the right. We stack boxes in the carts, but leave the big one shaped like a rocking horse. It's too heavy." "Why leave the big one?" "Because a rocking horse is not a present. It's a hostage. If we get cornered, we threaten to sit on it." Molly turned the corner. "Any questions?" "Yeah. Why are we stealing presents from children?" Molly paused. "Because adults have alarms. Children have distracted parents and a piñata. Now stop moralizing and pass me a juice box." Bill passed her the juice box. "What about the birthday cake? Do we take that too?" Molly shook her head. "Never take a cake. Frosting holds fingerprints." Bill considered this. "You've thought about this a lot." "I've thought about everything except your clown obsession." Molly sighed. "That's why I drive."
created_at
"2026-09-03T09:59:25.104Z"
9/3/2026, 9:59:25 AM
model
"deepseek/deepseek-v4-pro-20260813"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
984
moderation_latency
(null)
generation_time
45051
tokens_prompt
75
tokens_completion
3184
native_tokens_prompt
146
native_tokens_completion
3016
native_tokens_completion_images
(null)
native_tokens_reasoning
2364
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
0.01213608
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788429565-DnU0ANBDGHSlfYvhWjVt"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788429565-dG4UwHVbyg19Sq8aRfQk"
upstream_id
"chatcmpl-1a6117da-618c-4c11-beef-4b951f694e43"
provider_responses
0
endpoint_id
"65367950-3b6c-4abd-9cb0-d553a47de6d9"
id
"chatcmpl-1a6117da-618c-4c11-beef-4b951f694e43"
is_byok
false
latency
984
model_permaslug
"deepseek/deepseek-v4-pro-20260813"
provider_name
"SiliconFlow"
status
200
total_cost
0.01213608
cache_discount
(null)
upstream_inference_cost
0
provider_name
"SiliconFlow"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
3.62%
Matches word count
n/a
neededClean
false
words
476
0.00%
Dialogue to Total Word Ratio
Ratio: 79.25%, Deviation: 29.25%
neededClean
false
wordsTotal
477
wordsDialogue
378
1.8117%