NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-10-20-01M5C690NS9CMR51V37X1T1PJM
Run Details
Model:
DeepSeek V4.1 Flash 1017 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
47.2s
Cost:
$0.004338
Provider:
InferenceNet
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
15.29%
Dialogue to Total Word Ratio
(details)
7.6429%
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
The café was empty except for a sleeping waiter and a radio playing weather reports. Molly unfolded a crayon map across the sticky table while Bill moved the salt shaker onto the corner to keep it from curling. Bill studied the map upside down, which made the bouncy castle look like a wedding cake. “Saturday, two o’clock,” she said. “Lord Pemberton’s seventh birthday. Thirty kids, one magician, two exhausted parents, and a gift table guarded by a bouncy castle.” Bill frowned. “A bouncy castle has no arms.” “That’s what you said about the aquarium laser grid.” “We both got wet.” “Details. We go in as entertainment.” “Entertainment. Last time you entertained, the fire brigade came.” “That was a misunderstanding involving a unicycle. This is different. You’re a clown, I’m a fairy princess, and we’re early.” Bill wrote clown on a napkin and crossed it out. The napkin tore under the pressure. “Clowns are suspicious.” “Everybody trusts clowns.” “Exactly. What about a magician’s assistant?” “The magician brought doves. We can’t compete with doves.” “Fine. I’ll be a birthday bear.” “Too hot. You’ll pass out by cake.” “I’ll pass out after cake. Cake is the distraction.” Molly tapped the map. “No, the magician is the distraction. He does the handkerchief-and-egg thing. Parents watch. Kids watch. We collect every present and drop them into the bouncy castle.” Bill stared. “We steal the presents by putting them in a bouncy castle?” “Then we deflate it.” “You can’t deflate a bouncy castle at a party. It’s not a suitcase.” “It’s a suitcase with air. We pop one seam, it collapses, we drag it to the van.” “The kids will scream.” “Kids scream at balloons, cake, and the letter Q. They’ll think it’s part of the show.” “What about the auntie with the camera?” “That’s why we need the piñata.” “We’re not hitting a piñata.” “We don’t hit it. We tell the kids it’s full of money and let them riot. The auntie films that. Meanwhile, we—” “Steal the presents. Molly, this is insane.” “Insane is a strong word. I prefer festive.” “And the getaway?” “Party bags.” Bill blinked. The waiter snorted in his sleep and muttered about balloons. “Party bags?” “We fill two with the really good stuff, then walk out beside the parents, looking tired and holding cake.” “That’s not a getaway. That’s a stroll.” “Nobody runs from a birthday party unless there’s a clown.” Bill looked at his crossed-out napkin. “So I’m the clown.” “You’re the clown.” “I hate that this makes sense.” “Good. Practice your balloon giraffe. And don’t pop it.” “What if the snake pops?” “Then we improvise.” Bill folded the napkin into something that was definitely not a giraffe. “I can only do a snake.” “Then we’re stealing a snake’s birthday presents.”
created_at
"2026-10-20T08:19:04.511Z"
10/20/2026, 8:19:04 AM
model
"deepseek/deepseek-v4.1-flash-20261017"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
121
moderation_latency
(null)
generation_time
47171
tokens_prompt
75
tokens_completion
7662
native_tokens_prompt
67
native_tokens_completion
8665
native_tokens_completion_images
(null)
native_tokens_reasoning
7957
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
0.00433786
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1792484344-7cgOJzzPOe0CM3PVrqJa"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1792484344-w6K52webO0vEgaos7hmK"
upstream_id
"gv2_SbnLuniFF1pmsL9OMPcOp"
provider_responses
0
endpoint_id
"4231aea9-73af-4a2f-bd14-660b1d8b13b2"
id
"gv2_SbnLuniFF1pmsL9OMPcOp"
is_byok
false
latency
121
model_permaslug
"deepseek/deepseek-v4.1-flash-20261017"
provider_name
"InferenceNet"
status
200
total_cost
0.00433786
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
0.00%
Matches word count
n/a
neededClean
false
words
462
15.29%
Dialogue to Total Word Ratio
Ratio: 76.13%, Deviation: 6.13%
neededClean
false
wordsTotal
465
wordsDialogue
354
7.6429%