NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-10-20-01M5C6D6FX1PHW0EGEBTE8JZDT
Run Details
Model:
DeepSeek V4.1 Flash 1017 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
1m 16s
Cost:
$0.006641
Provider:
InferenceNet
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
65.78%
Dialogue to Total Word Ratio
(details)
82.8910%
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 crouched behind the van's steering wheel, peering through a crack in the windshield at the bouncing castle. "The clown's our ticket in. He's got a whistle and zero peripheral vision." Bill adjusted his fake mustache, which had slipped sideways during the drive, and checked the duffel bag of decoy bricks. "Those bricks better look like Lego sets. Last time you wrapped them in newspaper." Molly rolled her eyes and pointed at a cluster of parents hovering near the cake table like secret service agents. "That aunt's a drill sergeant. She pats down every gift bag for contraband juice boxes." Bill groaned and pulled a dog toy from his pocket, the squeaker already loose and rattling. "Swap the whistle with this. The clown will honk, not tweet. Parents will think it's comedy." Molly nodded and sketched a rough timeline on the dashboard with a grease pencil. "Two-fifteen, cake cutting. All eyes on frosting. We slide under the bounce house and puncture it slow." Bill stared at a toddler waddling past with a juice box and a suspiciously red stain on his shirt. "That kid's a biter. I still have scars from the last party. We need backup." Molly tossed him a juice box from the glove compartment, its straw still wrapped in plastic. "Backup is juice boxes. Toddler currency. Now focus on the gift table." Bill squinted through binoculars at a stack of presents wobbling beside the bouncy castle's inflation fan. "Lego sets, scooters, that fancy doll with the plastic crown. We swap them for bricks." Molly checked her watch, a cheap digital thing that beeped every hour like a tiny alarm. "Three minutes. Trip the clown after he drops a club. I'll grab the presents and run." Bill adjusted his magician's cape and practiced pulling a rabbit from a hat, but the rabbit was a stuffed sock. "If this goes wrong, I'm blaming the clown. He's got a suspicious nose for rubber chickens." Molly started the engine, which coughed twice before settling into a nervous idle. "Put on your mask. We're entertainers. Smile like you mean it. No biting." Bill pulled the ski mask over his face, leaving only his eyes and a twitching mustache visible. "What about the party bags? Every kid gets one. We can't take those." Molly grinned and revved the engine, scattering a flock of pigeons from the curb. "We don't. We take the big presents. Juice boxes for the crying kids. Done." Molly slowly pulled the old van into a side alley and killed the lights, leaving the tiny party glowing in the rearview mirror. "Just remember, no running. Little kids are short but surprisingly fast. And they bite hard. Always aim for their ankles first, okay please." Bill cracked nervously the van door and slipped onto the old pavement, then his magician's cape snagging on a loose rusty bolt again. "If anyone asks, we're the official entertainment. You're the magician. I'm the assistant. And the squeaky dog toy is very much a prop."
created_at
"2026-10-20T08:21:21.542Z"
10/20/2026, 8:21:21 AM
model
"deepseek/deepseek-v4.1-flash-20261017"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
164
moderation_latency
(null)
generation_time
76374
tokens_prompt
75
tokens_completion
11874
native_tokens_prompt
67
native_tokens_completion
13272
native_tokens_completion_images
(null)
native_tokens_reasoning
12606
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
0.00664136
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1792484481-sFGXbtvn2Ek3D7U5UOdB"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1792484481-sbcgcBOdPYI8MSpbAVd4"
upstream_id
"gv2_prDC84PTrvwmodgJGIXHV"
provider_responses
0
endpoint_id
"4231aea9-73af-4a2f-bd14-660b1d8b13b2"
id
"gv2_prDC84PTrvwmodgJGIXHV"
is_byok
false
latency
164
model_permaslug
"deepseek/deepseek-v4.1-flash-20261017"
provider_name
"InferenceNet"
status
200
total_cost
0.00664136
cache_discount
(null)
upstream_inference_cost
0
provider_name
"InferenceNet"
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
500
65.78%
Dialogue to Total Word Ratio
Ratio: 47.11%, Deviation: 2.89%
neededClean
false
wordsTotal
501
wordsDialogue
236
82.8910%