NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1KC5M7DKT55PWBP59JZK3V5
Run Details
Model:
DeepSeek V4 Flash 0731 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
30.6s
Cost:
$0.000192
Provider:
Relace
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 50% 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 shed smelled like lawnmower gas and regret. Bill was kneeling beside a duffel bag, holding up a rubber chicken like it was a sacred relic. Molly stood over him, arms crossed, wearing a balaclava that made her look like a very suspicious grape. “Explain the chicken,” Molly said. “It’s the distraction,” Bill said. “The distraction is a rubber chicken?” “Not just any rubber chicken. This one squeaks in the key of C. I tuned it myself.” “You tuned a rubber chicken.” “With a tiny little tuning fork. It took three hours.” Molly pinched the bridge of her nose. “Bill. We are breaking into a birthday party. For a six-year-old. There will be balloons, cake, and approximately forty presents. We need a plan, not poultry.” “The plan is the chicken,” Bill said, and he squeezed it. The squeak was unnervingly melodic. “No,” Molly said. “The plan is we walk in, I tell the parents I’m the balloon artist, you tell them you’re the face painter. While they’re distracted, we stuff the presents into these.” She kicked the duffel bag. “Then we leave through the hedge.” Bill stared at the chicken. “But what about the kids?” “What about them?” “They’ll see us,” Bill said. “Kids are like tiny security cameras with sticky fingers.” “That’s why we have the helium tank,” Molly said. “We fill a hundred balloons. Kids love balloons. They’ll mob us, we hand them out, and while they’re fighting over the red ones, we do the job.” Bill considered this. Then he shook his head. “Too complicated. Here’s my version. I walk in, I squeeze the chicken, everyone looks at me, and you grab the presents.” “Everyone looks at you because you’re a grown man squeaking a rubber chicken at a child’s party. They don’t just look. They call the police.” “It’s a funny noise, Molly. People laugh.” “The parents will not laugh when their kid’s presents are gone.” “They might not even notice,” Bill said. “Kids get a lot of presents. It’s chaos. You grab a few, no one counts.” “They count,” Molly said. “Parents always count. They have a mental inventory. ‘Oh, Grandma gave the Lego set, Aunt Carol gave the art supplies.’ They know.” “Okay, so we replace them with decoy presents,” Bill said, brightening. “We wrap bricks.” Molly closed her eyes. “We are not swapping a child’s birthday presents for bricks.” “Why not? It’s the perfect swap. We walk in with gift bags full of bricks, we put them under the table, we take the real ones, we leave. Nobody knows until tomorrow.” “They know today, Bill. The kid opens the bricks today. There is a weeping child covered in brick dust.” “I’ll wrap them nice,” Bill offered. Molly grabbed the rubber chicken out of his hand and threw it across the shed. It hit the wall and squeaked a sad, flat B-flat. “We do this my way,” she said. “Helium tank. Balloons. Duffle bag. Hedge. You do not make eye contact with any child, you do not speak unless spoken to, and you absolutely do not squeak anything.” Bill sighed. “Fine. But if the plan goes wrong, I’m going with the chicken.” “The plan won’t go wrong.” “That’s what you said about the bank.” “We robbed a bank, Bill.” “We robbed a branch of a savings and loan. They had a clown. The clown recognized me.” Molly stared at him. “The clown recognized you?” “We went to the same clown convention. He knew my stage name.” “You have a stage name?” “Squeaky Bill.” Molly turned and walked out of the shed. “We go at three. Wear a collared shirt.” “Squeaky Bill is not afraid of collared shirts!” Bill called after her, and he picked the rubber chicken off the floor, gave it a gentle, mournful squeeze, and started planning his own heist.
created_at
"2026-09-03T10:13:58.648Z"
9/3/2026, 10:13:58 AM
model
"deepseek/deepseek-v4-flash-20260731"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1089
moderation_latency
(null)
generation_time
20443
tokens_prompt
75
tokens_completion
1131
native_tokens_prompt
146
native_tokens_completion
1016
native_tokens_completion_images
(null)
native_tokens_reasoning
119
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
0.00019237
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788430438-B4eU2nclGqgf38a8NgqA"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788430438-XgiITXn5dyUTexTLrwMv"
upstream_id
"c4ac152633b044f1a07a8b75e8e9324d"
provider_responses
0
endpoint_id
"562d5188-97d1-49f7-8881-d0e417e31ada"
is_byok
false
latency
42
model_permaslug
"deepseek/deepseek-v4-flash-20260731"
provider_name
"Makora"
status
429
1
endpoint_id
"682c6334-ee4e-44f9-80d3-481aa5fd22b6"
is_byok
false
latency
222
model_permaslug
"deepseek/deepseek-v4-flash-20260731"
provider_name
"Ambient"
status
504
2
endpoint_id
"57c1bfab-049c-4d6a-ab34-ac1007a6043b"
id
"c4ac152633b044f1a07a8b75e8e9324d"
is_byok
false
latency
1089
model_permaslug
"deepseek/deepseek-v4-flash-20260731"
provider_name
"Relace"
status
200
total_cost
0.00019237
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Relace"
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
638
0.00%
Dialogue to Total Word Ratio
Ratio: 73.59%, Deviation: 23.59%
neededClean
false
wordsTotal
640
wordsDialogue
471
0.0000%