NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-10-20-01M5C6DKXRGH14Y2Y98N44SFK0
Run Details
Model:
DeepSeek V4.1 Flash 1017 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
1m 6s
Cost:
$0.005536
Provider:
InferenceNet
Evaluation summary
Score
Evaluator
Details Link
60.28%
Matches word count
(details)
90.15%
Dialogue to Total Word Ratio
(details)
75.2111%
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 unfolded the birthday invitation on the diner table, anchoring it with a saltshaker, a ketchup bottle, and a cold onion ring. The paper promised cake, games, and a bouncy castle, and someone had drawn a rocket ship in purple crayon. Bill studied the address, drew a square on his napkin, and checked the diner door twice before speaking. "The community center, Saturday at two. Presents in the blue room." "Correct. Window, no alarm, and a clown with a suspicious watch." "We go as Party Magic Plus. You balloon, I bag." Bill's eyebrow twitched. He had been banned from three magic shops and one petting zoo, but he still kept a collapsible wand in his sock. "The piñata is too small. We use the inflatable castle. Hidden zipper, false floor." Molly pulled a folded diagram from her sleeve and flattened it beside the ketchup. It showed a bouncy castle with dotted lines, arrows, and a small drawing of Bill dressed as a wizard. "You want to replace a child's bouncy castle with a bag." "Children love bags." Bill opened his mouth, closed it, and wrote the word liability on his napkin. "Last time you said children love smoke, we set off the alarm and a clown chased us with a balloon sword." "That clown was unstable. This is clean. We arrive at one forty-five. You do twenty minutes of magic. At two ten, the dove. At two twelve, I load presents. At two fifteen, we're gone." He drew little arrows and a stick figure dove with an angry expression. "What about the birthday boy?" "He'll be distracted by the dove." "That's not distraction. That's trauma with frosting." "Bill, focus. Parents watch the dove. The clown watches the petting zoo tip I gave him. Presents watch gravity." He folded his napkin into a tiny hat. "What if the clown talks?" "I told him there's a bigger score at the petting zoo." Molly tapped the diagram, which now had a tiny clown marked with an X. "There isn't." "Give me a week." Outside, a bus hissed and a child screamed happily at a balloon. Bill shuddered, then nodded. Bill sighed and checked the invitation again. "Fine. But I'm not wearing the cape." "You have to. It hides the rope." "What rope?" "The rope tying the bag to the handcart in the alley." "We have a handcart?" "I borrowed one from a very understanding man." She produced a key from her pocket and spun it on the table. "Understanding or unconscious?" "Flexible." Bill massaged his temples and reconsidered his career as a thief, a magician, and a man who had once been outsmarted by a piñata. Bill placed the napkin hat on the saltshaker. "One more thing. If a child asks me to pull a rabbit out of my hat, what do I do?" Molly smiled. "Pull out a parking ticket and tell them it's a magic car."
created_at
"2026-10-20T08:21:35.293Z"
10/20/2026, 8:21:35 AM
model
"deepseek/deepseek-v4.1-flash-20261017"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
119
moderation_latency
(null)
generation_time
66427
tokens_prompt
75
tokens_completion
9720
native_tokens_prompt
67
native_tokens_completion
11061
native_tokens_completion_images
(null)
native_tokens_reasoning
10413
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
0.00553586
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1792484495-WsqAq0x2zmJM5lViy3ap"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1792484495-8YWXnYQSJoJLnjkOWdJt"
upstream_id
"gv2_Jl1a52guRJpgNktQJ51qm"
provider_responses
0
endpoint_id
"4231aea9-73af-4a2f-bd14-660b1d8b13b2"
id
"gv2_Jl1a52guRJpgNktQJ51qm"
is_byok
false
latency
119
model_permaslug
"deepseek/deepseek-v4.1-flash-20261017"
provider_name
"InferenceNet"
status
200
total_cost
0.00553586
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
60.28%
Matches word count
n/a
neededClean
false
words
485
90.15%
Dialogue to Total Word Ratio
Ratio: 51.44%, Deviation: 1.44%
neededClean
false
wordsTotal
486
wordsDialogue
250
75.2111%