NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1KBDTR8GESCM9ANF1K8RXTG
Run Details
Model:
DeepSeek V4 Pro 0813 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
22.1s
Cost:
$0.014597
Provider:
BaseTen
Evaluation summary
Score
Evaluator
Details Link
99.98%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
49.9920%
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 leaned over the crumpled napkin map. "Okay, Bill, the bounce house is the east entrance. You go in low, past the juice boxes." Bill squinted. "Why low? I'm six-two. Low means I'm basically crawling through toddlers." "Exactly. They'll think you're a dad playing horsey. You'll blend." "I don't blend with glitter and cupcake frosting, Molly. Last time I wore a party hat, I got glitter in my eye for a week." "That's the price of a score this big. Now, the gift table is by the pinata. You distract the clown, I'll load the wagon." "What wagon?" "The little red one with the handle. It's already there. I scouted it." Bill sighed. "You want me to distract a clown. How? Balloon animals? A pie to the face?" "Tell him you're the entertainment coordinator and his check bounced. That'll buy us four minutes." "Four minutes? You're going to steal forty presents in four minutes?" "No, I'm going to steal forty presents in three. The fourth minute is for the chocolate fountain." Bill rubbed his temples. "Molly, it's a six-year-old's birthday party. There will be parents. Parents who notice a strange woman shoving gift bags into a wagon." "Not if I'm wearing the mascot costume." "What mascot?" "Barry the Birthday Beaver. The zipper sticks, but the head is roomy." Bill stared. "You want me to distract a clown while you, dressed as a beaver, rob a child's gift table. And I'm supposed to be the weird one?" "Bill, focus. The clown is the linchpin. If he cries, the kids cry, and crying kids draw parents. We need him angry, not sad." "So I insult his check. Then what?" "Then you slip out through the face-painting station. The face painter owes me a favor. She'll give you a full tiger face. You'll walk out looking like a hired entertainer." "A tiger face. Great. That'll help me carry the wagon." "No, I'm carrying the wagon. You're the lookout. You stand by the bouncy castle and shout 'the clown is doing magic' if anyone gets near me." "What if the clown follows me?" "Then you challenge him to a balloon sword fight. You'll lose, but it'll be funny." Bill crossed his arms. "I'm not fighting a clown with a balloon sword." "Fine. Tell him there's a second bounce house out back with a broken waiver. He'll sprint." Molly tapped the napkin. "We meet at the minivan at two-fifteen. The party ends at two-thirty. We'll be gone before the cake." Bill shook his head. "This is the dumbest plan we've ever had." Molly grinned. "That's what you said about the petting zoo heist, and we made four hundred bucks." "We made four hundred bucks because a goat ate the evidence." "Exactly. Trust the chaos, Bill." Bill sighed and pocketed the napkin. "Fine. But if I get glitter in my eye again, I'm keeping the biggest present." "Deal. Now practice your clown face. You look too happy." "That's my terrified face, Molly." "Perfect. Terrified clowns are very believable."
created_at
"2026-09-03T10:00:58.901Z"
9/3/2026, 10:00:58 AM
model
"deepseek/deepseek-v4-pro-20260813"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
628
moderation_latency
(null)
generation_time
21713
tokens_prompt
75
tokens_completion
3412
native_tokens_prompt
147
native_tokens_completion
3637
native_tokens_completion_images
(null)
native_tokens_reasoning
2926
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
0.01459656
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788429658-cPjPASshqS8D0DOzbFfV"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788429658-iXIsHLa55ntvnx5HRSJs"
upstream_id
"chatcmpl-2d5b13bed7304044977aad9c1475bef4"
provider_responses
0
endpoint_id
"6811c06b-8624-4796-ad6e-807e33b97ae6"
is_byok
false
latency
300
model_permaslug
"deepseek/deepseek-v4-pro-20260813"
provider_name
"Sail Research"
status
429
1
endpoint_id
"85bb5a4a-983d-4d11-b97b-c7e5b348375f"
id
"chatcmpl-2d5b13bed7304044977aad9c1475bef4"
is_byok
false
latency
628
model_permaslug
"deepseek/deepseek-v4-pro-20260813"
provider_name
"BaseTen"
status
200
total_cost
0.01459656
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
99.98%
Matches word count
n/a
neededClean
false
words
502
0.00%
Dialogue to Total Word Ratio
Ratio: 92.72%, Deviation: 22.72%
neededClean
false
wordsTotal
508
wordsDialogue
471
49.9920%