NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-10-20-01M5C6H5BSZ0FBKSVQ8A4291NJ
Run Details
Model:
DeepSeek V4.1 Flash 1017 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
44.7s
Cost:
$0.004055
Provider:
InferenceNet
Evaluation summary
Score
Evaluator
Details Link
6.09%
Matches word count
(details)
99.88%
Dialogue to Total Word Ratio
(details)
52.9875%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 70% 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 parked the van behind a bouncy castle shaped like a dragon and cut the engine. The party hall windows throbbed with a song about sharing. Bill unfolded a napkin covered in crayon arrows and held it at arm’s length, as if it might explode. Through the window, a pyramid of wrapped boxes glittered under a banner that read HAPPY BIRTHDAY, TIMMY. A child in a superhero cape was licking the cake. A hired clown honked a horn and looked professionally miserable. Molly counted six exits, four of them blocked by strollers. Bill folded the napkin into a hat. It collapsed. Molly frowned with visible disappointment. “Walk me through it again, Bill. Slowly. Last time you said simple, we ended up with a goat in the getaway car.” “The party starts at two fifteen. Twenty-seven kids, one clown, two prosecco mums pretending to supervise. We go in as entertainment.” “As what? You look like a debt collector who lost a fight with a piñata.” “I’m the magician. You’re the assistant. We wheel in the big box, I say abracadabra, and while I pull a rabbit out of a hat, you load every present into the box.” Molly stared at the napkin. “Every present?” “Every present. The table by the cake, the mountain by the coat rack, and the suspicious pile the birthday boy guards like a dragon with a trust fund.” “And the clown?” “Wild card. He does balloon animals and judges everyone. If he asks questions, distract him with a balloon sword.” “You want me to fight a clown with a balloon sword?” “I want you to look confident. It’s eighty percent confidence, fifteen percent timing, five percent not getting glitter on the evidence.” Molly lowered the window. The party hall smelled of icing and expensive shoes. “That math feels generous.” “Also, the bouncy castle is our exit. We cut the tie-downs, the castle rolls toward the fence, we surf it into the alley, presents secured.” “That’s not surfing. That’s being crushed by a moon bounce in front of toddlers.” “Details.” “Bill, we are stealing from a children’s party.” “We’re liberating excess consumer goods from a ritual of enforced generosity.” Bill tapped the napkin. “The flyer says bring a gift, leave with a smile. We’re just leaving with more gifts than smiles.” “And if the kids see us?” “Kids love chaos. We tell them it’s a game called Find the Presents Later.” “That is the worst game ever.” “It has stakes.” Molly checked the wing mirror. A clown was already tying balloon poodles to the fence. “What about the parents?” “They’ll assume we’re hired. Nobody questions a man in a cape carrying a box.” “You’re not wearing a cape.” “I’ve got a tablecloth in the back with ambition.” “And the birthday boy? The dragon with a trust fund?” “He’s four. We give him a juice box and a fake moustache. He’ll forget his own name.” Molly sighed and opened the glovebox. “Fine. But if I have to sword-fight a clown, you’re explaining it to the judge.” Bill grinned and reached for the sparkly waistcoat. “Deal. We’ve got a party to ruin tastefully.”
created_at
"2026-10-20T08:23:31.454Z"
10/20/2026, 8:23:31 AM
model
"deepseek/deepseek-v4.1-flash-20261017"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
214
moderation_latency
(null)
generation_time
44673
tokens_prompt
75
tokens_completion
7290
native_tokens_prompt
67
native_tokens_completion
8099
native_tokens_completion_images
(null)
native_tokens_reasoning
7367
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
0.00405486
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1792484611-WOv9QKb5bdC2GrIqDtz7"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1792484611-hcsDUp6A1Ja2OCCoD1bw"
upstream_id
"gv2_w1WJodWSG3973LENPBGo5"
provider_responses
0
endpoint_id
"4231aea9-73af-4a2f-bd14-660b1d8b13b2"
id
"gv2_w1WJodWSG3973LENPBGo5"
is_byok
false
latency
214
model_permaslug
"deepseek/deepseek-v4.1-flash-20261017"
provider_name
"InferenceNet"
status
200
total_cost
0.00405486
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
6.09%
Matches word count
n/a
neededClean
false
words
523
99.88%
Dialogue to Total Word Ratio
Ratio: 70.15%, Deviation: 0.15%
neededClean
false
wordsTotal
526
wordsDialogue
369
52.9875%