NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-16-01KHM5SRHQZ4BGV5YTSSZ7X7MF
Run Details
Model:
DeepSeek V3.2
Test:
Dialogue tags
Scenario:
Write 200 words with 50% dialogue
Duration:
7.4s
Cost:
$0.000131
Provider:
Novita
Model input
User/Human
You are a author of comedic heist books. Please write a 200 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
The warehouse loft smelled of dust and distant rain. Molly traced a finger over the blueprints, her expression grim. “It’s a hard target, Bill. Perimeter is a backyard fence. Guard rotation is two parents, likely sleep-deprived. Primary threat: twenty sugar-fueled four-year-olds.” Bill adjusted his black beanie. “So we go in quiet. I’ll create a diversion with a rogue piñata. You sweep the gift table.” “The payload is wrapped in loud, crinkly paper. Extracting it without alerting the tiny, screaming sentries is the issue.” “We’ll need a transport. Big red bag, maybe. Blends with the theme.” Molly snorted. “Santa’s gone bad. I don’t like it. The intel says the big gift is a ‘Talking Princess Celestial Unicorn.’ High-value, but it probably glows and sings.” “It’s the principle, Moll. The job’s the job.” Bill pointed at the schematic. “We hit during ‘Musical Chairs.’ Maximum chaos.” “And the exit?” “Same way we came in. Over the fence, through old Mrs. Henderson’s petunias.” Molly sighed, a sound of pure professional resignation. “Fine. But if I get glitter on me, I’m leaving you behind.”
created_at
"2026-02-16T21:30:52.126354+00:00"
2/16/2026, 9:30:52 PM
model
"deepseek/deepseek-v3.2-20251201"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
994
moderation_latency
(null)
generation_time
7406
tokens_prompt
77
tokens_completion
315
native_tokens_prompt
67
native_tokens_completion
282
native_tokens_completion_images
(null)
native_tokens_reasoning
0
native_tokens_cached
0
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
0.000130823
router
(null)
provider_responses
0
id
"df367aa1586458dd869d2ecd047166ab"
status
200
is_byok
false
latency
994
endpoint_id
"8a5fb6d3-ef71-4958-af54-269ffa3a3c6e"
provider_name
"Novita"
model_permaslug
"deepseek/deepseek-v3.2-20251201"
api_type
"completions"
id
"gen-1771277443-dEUlRGKVgm2hceLZ7lUE"
upstream_id
"df367aa1586458dd869d2ecd047166ab"
total_cost
0.000130823
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Novita"
Evaluation details
Result
Evaluator
Details
Meta Data
51.9255%
Matches word count
n/a
neededClean
false
words
184
0.0000%
Dialogue to Total Word Ratio
Ratio: 78.80%, Deviation: 28.80%
neededClean
false
wordsTotal
184
wordsDialogue
145
25.9628%