NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-10-20-01M5C6GW93ATY5H6BGE48JHT0F
Run Details
Model:
DeepSeek V4.1 Flash 1017 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
53.2s
Cost:
$0.004529
Provider:
InferenceNet
Evaluation summary
Score
Evaluator
Details Link
1.04%
Matches word count
(details)
50.73%
Dialogue to Total Word Ratio
(details)
25.8826%
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 spread the birthday flyer across the café table, anchoring one corner under a ketchup bottle and the other beneath a cold toast crust. Rain slid down the window while she tapped the community hall’s address with a chipped fingernail. Bill leaned in, pencil behind his ear, and squinted at the pirate hat drawn in the corner. The café smelled of burnt espresso and wet umbrellas. Bill opened a battered notebook labeled Party Ideas, which was also his fencing ledger. Molly ignored the health inspection certificate curling above the sugar dispenser and drew a small square for the present table. “Target?” Bill asked. “Seven-year-old’s pirate party,” Molly said. “Forty kids, one bouncy castle, and a present table groaning like a galleon.” Bill wrote “galleon” on a napkin. “We go as pirates, then.” “No cutlasses. Last time you nearly severed a waiter’s ear over a breadstick.” “That breadstick was asking for it.” Molly took the pencil and drew a dotted line from the kitchen door to the fire exit. A waitress dropped off two teas and a plate of biscuits shaped like tiny lighthouses, then lingered until Bill hid the napkin under his elbow. Molly rotated the flyer. “We need access and exit. Kitchen door, fire exit, front entrance blocked by a clown.” Bill shuddered. “I hate clowns. They smell like balloons.” “You’ll survive. You’re a magician; I’m a face painter.” “I can’t do magic.” “You can look disappointed in a top hat. Kids love that.” Bill scratched his chin. “Presents? They’ll be wrapped. We won’t know which are good.” “We take all of them. It’s a heist, not a book club.” “Some are socks.” “Socks go to the shelter. We’re thieves, not monsters.” Bill nodded slowly. “Distraction?” Molly smiled. “The piñata. You hit it too hard, sweets explode, chaos, we slip out with three sacks.” “Three? I have two arms.” “Use your back.” “My back’s busy standing upright.” Molly sighed. “Fine. A wheelbarrow painted like a treasure chest.” “Suspicious.” “It’s a pirate party, Bill. Suspicious is the theme.” Bill sketched the hall on his napkin. “Parents will notice a wheelbarrow rolling out.” “That’s why we hire a decoy clown.” “You hate clowns.” “In the abstract. For business, I respect them.” Bill drew a tiny clown, then crossed it out. The café clock clicked toward noon. Molly folded the flyer into a neat square and slid it under the ketchup bottle, leaving one corner visible like a dare. “What about the parents?” “You tell a boring story about maritime law. They’ll drift off.” “I don’t know maritime law.” “Perfect. It’ll be so boring they’ll leave. That’s the heist.” “What if a kid sees us?” “We say we’re elf-pirates.” Bill stared at the napkin, then at Molly. “This is the stupidest plan you’ve ever had.” “Thank you.” “When do we start?” “Saturday. Comfortable shoes. No cutlass.”
created_at
"2026-10-20T08:23:22.153Z"
10/20/2026, 8:23:22 AM
model
"deepseek/deepseek-v4.1-flash-20261017"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
138
moderation_latency
(null)
generation_time
53136
tokens_prompt
75
tokens_completion
8271
native_tokens_prompt
67
native_tokens_completion
9048
native_tokens_completion_images
(null)
native_tokens_reasoning
8328
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
0.00452936
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1792484602-07HFjW3LJEuSZiVXylWd"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1792484602-xiWvVXRuiOR0ijsaJTjc"
upstream_id
"gv2_ve3hgWIFSeRoj4ECN51sZ"
provider_responses
0
endpoint_id
"4231aea9-73af-4a2f-bd14-660b1d8b13b2"
id
"gv2_ve3hgWIFSeRoj4ECN51sZ"
is_byok
false
latency
137
model_permaslug
"deepseek/deepseek-v4.1-flash-20261017"
provider_name
"InferenceNet"
status
200
total_cost
0.00452936
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
1.04%
Matches word count
n/a
neededClean
false
words
474
50.73%
Dialogue to Total Word Ratio
Ratio: 53.68%, Deviation: 3.68%
neededClean
false
wordsTotal
475
wordsDialogue
255
25.8826%