NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1KBQ02S1H9MSFP6ZHGP1XWT
Run Details
Model:
DeepSeek V4 Pro 0813 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
30.8s
Cost:
$0.003755
Provider:
DigitalOcean
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
0.0004%
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 apartment smelled like burnt microwave popcorn and regret. Molly was hunched over the kitchen table, tracing a crayon drawing of a unicorn with a butter knife. Bill was trying to remove a wad of pink bubblegum from the sole of his loafer with a fork. “This is a new low, Mol.” “It’s not a low. It’s a diversification of assets.” “It’s a bouncy castle, Bill. I haven’t been in one of those since my cousin’s communion.” “Exactly. No one expects grown adults in there. It’s the perfect cover.” “Are we hiding in the bouncy castle?” “We might have to.” “I don’t know if my knees can handle that.” “Your knees will handle a thousand dollars in unopened Legos.” Bill stopped attacking the gum and stared. “A thousand dollars?” “Easily. This is Westport. These parents don’t buy little Tommy a twelve-dollar action figure. They buy a six-hundred-dollar ride-on tractor with a real cup holder.” “We could sell that for half.” “Quarter. But still.” Bill scratched his chin. “So we just walk in? ‘Hi, we’re here for the clown simulator’?” “No, we go as the entertainment.” “Which is what?” “A magician and his lovely assistant.” Bill blinked. “I’m not wearing sequins.” “You’re the magician. I’m the assistant. I’ll wear sequins. You just need a cape and a top hat.” “Where do we get a top hat?” “Party City. They don’t ask questions.” “They literally ask if you want to add balloons.” “So we add balloons. It’s a distraction. You’re overthinking this, Bill.” “I’m not overthinking. I’m under-thinking. That’s the problem.” Molly set down the butter knife. “Look. Here’s the play. We show up twenty minutes after the party starts. The parents are already exhausted. The kids are screaming. Nobody checks your ID.” “Except the birthday mom who hired a magician two months ago and knows we’re not him.” “So we say the original magician got salmonella. We’re the backup.” “And she just believes that?” “She will when I pull a rabbit out of the birthday girl’s ear.” “We don’t have a rabbit.” “We have a hamster. Same energy. Smaller ears.” Bill groaned and dropped the fork. “This plan has more holes than my last alibi.” “My last alibi kept you out of jail.” “That’s because the cop had allergies and couldn’t stop sneezing long enough to write me up.” “Preparation, Bill.” “That wasn’t preparation, that was luck and a ragweed bloom.” “Fine. What’s your big idea?” “Don’t go to the party. Wait until it’s over. Break into the house at two in the morning.” “And steal what? Leftover cake and a half-deflated balloon?” “And the presents. They leave them in the living room. Kids don’t open gifts until the next day.” Molly considered this. “That’s boring.” “Boring is profitable. Boring gets you out before the sun comes up.” “But we don’t get any cake.” “We can buy our own cake.” “Not with a thousand dollars in Legos, Bill. Then we can buy a yacht made of cake.” “That’s not a real thing.” “It is if you dream big.” Bill sighed and looked at the crayon drawing again. “If I agree to the bouncy castle, do I have to do balloon animals?” “Only if they tip.” “Kids don’t tip.” “Then just pull the hamster out and run for the gifts.” “This is the worst idea I’ve ever agreed to.” “That’s the spirit, partner.”
created_at
"2026-09-03T10:05:59.274Z"
9/3/2026, 10:05:59 AM
model
"deepseek/deepseek-v4-pro-20260813"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1405
moderation_latency
(null)
generation_time
30761
tokens_prompt
75
tokens_completion
1035
native_tokens_prompt
67
native_tokens_completion
926
native_tokens_completion_images
(null)
native_tokens_reasoning
76
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
0.0037554
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788429959-edkYwAWUPws0HfbTvSPp"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788429959-qYWGJ8ePKd2YhYLzs6uT"
upstream_id
"chatcmpl-fa596ec3-7e93-4f41-a7cc-d34b80d5704e"
provider_responses
0
endpoint_id
"928ab2e9-a85f-427f-adcb-15b2394c70d5"
id
"chatcmpl-fa596ec3-7e93-4f41-a7cc-d34b80d5704e"
is_byok
false
latency
1405
model_permaslug
"deepseek/deepseek-v4-pro-20260813"
provider_name
"DigitalOcean"
status
200
total_cost
0.0037554
cache_discount
(null)
upstream_inference_cost
0
provider_name
"DigitalOcean"
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
561
0.00%
Dialogue to Total Word Ratio
Ratio: 85.36%, Deviation: 15.36%
neededClean
false
wordsTotal
567
wordsDialogue
484
0.0004%