NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1KD0680PR52X84ZPNKKHV89
Run Details
Model:
DeepSeek V4 Flash 0731 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
11.9s
Cost:
$0.000273
Provider:
BaseTen
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
0.0000%
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 herself against the hedge, binoculars trained on the backyard. Bill crouched beside her, picking a leaf out of his ear. “I count forty-seven balloons,” she whispered. “That’s a lot of helium.” “Forty-seven balloons means a lot of parents,” Bill said. “And parents mean phones. And phones mean cops.” “But also presents. Stacked on that table. Look at them. All shiny. Wrapped in that crinkly paper that screams ‘I cost more than a sandwich.’” Bill squinted. “The princess one’s got a bow the size of a small dog.” “That’s the target. The princess one. Also the one shaped like a dinosaur. And the one that’s just a suspiciously large envelope.” “We’re not stealing an envelope.” “We’re stealing everything, Bill. That’s the plan. We go in as clowns.” “Clowns.” “Yes. We have the costumes. I’ve got the red nose. You’ve got the squirting flower.” “I hate the squirting flower. It’s demeaning.” “It’s a distraction. While the kids are screaming and the adults are trying to figure out if we’re hired or just tragic, we grab the haul and slip out the side gate.” “They won’t scream. They’ll just stare. Kids are weird about clowns. They go quiet. Then one of them says something like ‘I know you’re not a real clown because your shoes aren’t funny enough.’ That’s what happened last time.” “That was a different party. And you wore sandals.” “I had a blister.” “We don’t have time for your feet, Bill. The party starts at two. We need to be in position by one-forty-five. I’ve got the van parked behind the Dumpster. We’ll do the balloon-animal routine for exactly four minutes. Then I ‘accidentally’ knock over the punch bowl. You grab the presents. We run.” “What about the birthday kid?” “What about her?” “She’s going to be very upset.” “She’s going to get a lot of new presents next year. It’s a growth industry.” “You have a cold heart, Molly.” “I have a warm heart. It’s just also practical. Look, Bill, we’re not villains. We’re redistributing wealth from the over-indulged to the under-employed. That’s us. We’re the under-employed.” “We’re thieves.” “Semantics. Are you in or not?” Bill sighed. He pulled the wilted flower from his pocket and squirted himself in the face. “Fine. But I’m wearing my own shoes.” “Deal. Now, let’s go over the signal. When I say ‘Who wants a balloon?’ that means the punch bowl is about to meet the carpet. When I say ‘Who wants a balloon?’ in a higher pitch, that means I’ve spotted a parent with a camera. When I say ‘Who wants a balloon?’ in a low, growly voice, that means I’ve spotted a dog.” “What do I do if you see a dog?” “Grab the presents faster.” “Right.” “And Bill?” “Yeah?” “If you see a kid with a juice box, don’t make eye contact. They throw them. It’s a whole thing.” “I remember.” “Good. Now let’s go over the escape route one more time. We go out the side gate, left past the bouncy castle, right through the rose bushes, over the fence, and into the van. If anyone asks, we’re delivering a cake.” “There’s no cake in the van.” “There’s a cake in the van. I stole one this morning. It’s a nice one. Chocolate.” Bill nodded slowly. “You’re a professional, Molly. I respect that.” “Thank you. Now put on your nose. It’s showtime.”
created_at
"2026-09-03T10:28:29.079Z"
9/3/2026, 10:28:29 AM
model
"deepseek/deepseek-v4-flash-20260731"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
383
moderation_latency
(null)
generation_time
11831
tokens_prompt
75
tokens_completion
1055
native_tokens_prompt
147
native_tokens_completion
975
native_tokens_completion_images
(null)
native_tokens_reasoning
114
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
0.00027261
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788431309-dze9Xlh4h5tAa3X6RlRV"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788431309-TIxXGEPAxTO1DDgFFUp6"
upstream_id
"chatcmpl-45e271298638401b902819f0d5ce17bd"
provider_responses
0
endpoint_id
"ceaae883-4cba-45dd-b986-5cd0e6bf06af"
id
"chatcmpl-45e271298638401b902819f0d5ce17bd"
is_byok
false
latency
383
model_permaslug
"deepseek/deepseek-v4-flash-20260731"
provider_name
"BaseTen"
status
200
total_cost
0.00027261
cache_discount
(null)
upstream_inference_cost
0
provider_name
"BaseTen"
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
565
0.00%
Dialogue to Total Word Ratio
Ratio: 91.78%, Deviation: 21.78%
neededClean
false
wordsTotal
572
wordsDialogue
525
0.0000%