NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-10-20-01M5C6CYN2VEM2S6RHMH46EHY5
Run Details
Model:
DeepSeek V4.1 Flash 1017 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
1m 17s
Cost:
$0.006684
Provider:
InferenceNet
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
99.97%
Dialogue to Total Word Ratio
(details)
99.9836%
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 flattened the invitation on the café table and tapped the guest list like a general. “Right, Bill: community hall, two o’clock, presents on the trestle table by the cake. We take everything.” Bill considered this. “Even the wrapped socks?” “Especially the socks. Volume, Bill. Volume.” “I thought we were after the porcelain unicorn.” “That’s phase two. Phase one is the present avalanche.” Bill scratched his chin. “Do we dress as clowns?” “No. You dress as a clown. I’m the responsible parent who’s very sorry about you.” “It’s a role. You juggle, trip over the bouncy castle, and create a diversion near pass-the-parcel.” “What if the real clown objects?” “He won’t. You’re more chaotic. He’ll step aside out of professional respect.” “And the parents?” “Easy. They’re holding phones and half a slice of cake. I say I’m from the entertainment agency and ask if the magician arrived. While they panic, you go for the table.” Across the table, a waiter set down two coffees. Molly ignored hers and began arranging sugar packets into a scale model of the hall. “What about the kids?” “Kids are the problem. Excellent peripheral vision, no impulse control. We wait for the birthday song. Everyone looks at the cake. Nobody looks at the presents.” Bill wrote KIDS on his hand and underlined it twice. Then he underlined cake, because he liked cake. Molly drew a small arrow pointing at the word CAKE and another pointing at the word SACRIFICE. Bill nodded as if this clarified everything. “What do I do during the song?” “Sing. Loudly. Slightly wrong. Off-key enough that people look at you, not the table.” “Then I sweep the presents into the Santa sack. We exit through the kitchen. The door sticks, so lift, don’t push.” “What if someone follows?” “They won’t. You release the balloons.” “No, not balloon animals. The big helium ones. Twelve of them. Instant chaos.” Bill frowned. “Molly, are we stealing from children?” “We’re redistributing joy. Very gradually. Through a pawn shop.” “Fine. We’re stealing from children, but tastefully.” “That’s negotiating. Now, the getaway. I’ve got a hatchback outside with a blanket over the plates.” Bill produced a pair of oversized red shoes from under the table. They squeaked when he set them down. The shoes were too big for him and too small for the plan. Outside, a pigeon regarded the hatchback with what looked like suspicion. “We’re not taking the bouncy castle.” “Why not?” “Because it won’t fit, Bill.” “We could deflate it.” “I’m just saying it’s mostly air.” “So are you, Bill, but we’re not deflating you either.” “We are not deflating a child’s bouncy castle during the birthday song. There are limits.” Molly stared at him until he put the shoes on. The squeak was immediate and deeply unprofessional. “Bill, focus. Two o’clock. Clown shoes. Santa sack. Cake song. Balloons. Kitchen door. Hatchback.” “And the unicorn?” “If the unicorn survives the avalanche, yes. But it’s not the priority.” “Put on the clown shoes, Bill.”
created_at
"2026-10-20T08:21:13.512Z"
10/20/2026, 8:21:13 AM
model
"deepseek/deepseek-v4.1-flash-20261017"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
83
moderation_latency
(null)
generation_time
77323
tokens_prompt
75
tokens_completion
12344
native_tokens_prompt
67
native_tokens_completion
13358
native_tokens_completion_images
(null)
native_tokens_reasoning
12612
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
0.00668436
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1792484473-IWtUyWXMbx9fZBkZAFSS"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1792484473-oUycTwKjzvtKR1AelxHm"
upstream_id
"gv2_2ZZzoolmw2GdXolmv7KmW"
provider_responses
0
endpoint_id
"4231aea9-73af-4a2f-bd14-660b1d8b13b2"
id
"gv2_2ZZzoolmw2GdXolmv7KmW"
is_byok
false
latency
83
model_permaslug
"deepseek/deepseek-v4.1-flash-20261017"
provider_name
"InferenceNet"
status
200
total_cost
0.00668436
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
499
99.97%
Dialogue to Total Word Ratio
Ratio: 69.92%, Deviation: 0.08%
neededClean
false
wordsTotal
502
wordsDialogue
351
99.9836%