NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-16-01KHM5T03EECYH6JN3RHT8RTQG
Run Details
Model:
DeepSeek V3 (2024-12-26)
Test:
Dialogue tags
Scenario:
Write 200 words with 50% dialogue
Duration:
12.2s
Cost:
$0.000423
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
Molly adjusted her fake mustache in the rearview mirror while Bill fiddled with the oversized clown shoes in his lap. “You sure this is the right move?” Bill asked, squinting at the party invitations scattered on the dashboard. “Breaking into a six-year-old’s birthday feels… low.” Molly scoffed. “Low? Bill, it’s genius. Kids get the best loot these days—designer toys, gold-plated rattles, whatever. Plus, security’s a joke. One guy in a dinosaur costume and a piñata full of regrets.” Bill sighed. “What if we get caught? You ever tried outrunning cops in these shoes?” He wobbled one in the air. “Relax. We’ll blend in. You’re the clown, I’m the magician—” “You’re wearing a trench coat and a mustache.” “—*Disguised* as a magician. Look, all we do is grab the presents during the cake chaos, load ‘em into the balloon animal van, and boom. Retirement.” Bill hesitated. “What about the kids?” Molly rolled her eyes. “They’ll get over it. Childhood trauma builds character.” A beat. Then Bill shrugged. “Fine. But if I have to make one balloon poodle, I’m out.” Molly grinned. “Deal. Now let’s go rob some toddlers.”
created_at
"2026-02-16T21:31:04.312462+00:00"
2/16/2026, 9:31:04 PM
model
"deepseek/deepseek-chat-v3"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1431
moderation_latency
(null)
generation_time
11935
tokens_prompt
77
tokens_completion
331
native_tokens_prompt
66
native_tokens_completion
305
native_tokens_completion_images
0
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
0.0004229
router
(null)
provider_responses
0
id
"ccfbf3bd136be257347c3aba88ed8e86"
status
200
is_byok
false
latency
1431
endpoint_id
"cefb5a1e-9fea-4496-9a64-4e1b11a7cc8c"
provider_name
"Novita"
model_permaslug
"deepseek/deepseek-chat-v3"
api_type
"completions"
id
"gen-1771277451-J7v7CSckO6OEKHSPWM2M"
upstream_id
"ccfbf3bd136be257347c3aba88ed8e86"
total_cost
0.0004229
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Novita"
Evaluation details
Result
Evaluator
Details
Meta Data
93.6496%
Matches word count
n/a
neededClean
false
words
191
0.0000%
Dialogue to Total Word Ratio
Ratio: 71.73%, Deviation: 21.73%
neededClean
false
wordsTotal
191
wordsDialogue
137
46.8248%