NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-04-25-01KQ2DPVCZGZFDXSF1YTMAZAJ2
Run Details
Model:
DeepSeek V4 Flash
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
12.2s
Cost:
$0.000231
Provider:
DeepSeek
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 30% 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 van smelled like stale coffee and guilt. Bill was peering through a pair of binoculars, his face pressed against the grimy window. “Okay, I’ve got three piñatas. One is a dump truck, which is just lazy. One is a unicorn that looks like it’s seen things. And one is a… is that a taxidermy squirrel?” Molly lowered the map of the backyard. “Focus, Bill. The target is the gift table.” Bill grunted. “I’m focused. I’m doing reconnaissance. There’s a castle bounce house. A small Rottweiler in a party hat. And a man in a unicorn onesie who is currently crying near the grill.” “That’s the father,” Molly said. “He always cracks. We go in during the ‘Happy Birthday’ song. It’s the only time the adults aren’t looking at the loot.” “And what’s my job?” Bill asked, turning. He had a smear of chocolate on his cheek from his fourth donut. “You’re the distraction. We discussed this. The smoke machine is in the back. When the kids are coughing, you’ll wobble in wearing the gorilla costume.” Bill looked appalled. “The one with the butt zipper? The kids are going to see the zipper. They have eyes, Molly. Sharp, judgmental eyes.” “They’re five, Bill. Their main concern is whether the cake has a toy ring on it. Now, the plan. I enter from the hedge line, use the leaf blower to create a diversionary leaf-tornado, and while the magician is trying to comfort the weeping unicorn dad, I swap the real presents with these decoy boxes filled with rocks.” Bill scratched his head. “And what about the pony ride? There’s a Shetland pony. It’s tied to a tree. It looked at me with disappointment.” “Ignore the pony. Its judgment is irrelevant. Your job is to start a conga line with the gorilla head on backwards. Make it look like a tragic accident. The kids will find it hilarious. The adults will be confused. I get a clean thirty seconds.” “Thirty seconds for how many gifts?” Bill asked, his voice a low whine. “I counted twelve. Wrapped in that glittery paper that sheds. It gets everywhere.” “It gets in the decoy boxes too. That’s the realism, Bill.” A child’s scream pierced the air. A beat. Then the tinny recording of “Happy Birthday” began, crackling from a cheap speaker. Bill’s eyes went wide. “They’re starting. I can’t. I’m having a panic attack. My donuts are rebelling.” Molly grabbed his shoulders. “Listen to me. You are not a failed clown. You are a gorilla of opportunity. You are going to walk into that backyard, make a fool of yourself, and when you hear the scream about the pony escaping, you know I’m en route.” “The pony escapes?” “In the plan, yes. I’ve got a carrot and a laser pointer. Now, go. Be a nightmare. A ridiculous, hairy, zipper-butted nightmare.” Bill sighed, a sound of pure resignation. “If I get glitter in my fur, I’m naming my firstborn after that pony.” He clambered out the back of the van, the gorilla suit squeaking with each step. Molly watched him shuffle toward the fence, a lonely, shaggy silhouette against the bobbing helium balloons. She checked the leaf blower, hefted the box of rocks, and whispered to herself, “The gift of regret. For everyone.”
created_at
"2026-04-25T13:36:48.039Z"
4/25/2026, 1:36:48 PM
model
"deepseek/deepseek-v4-flash-20260423"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
720
moderation_latency
(null)
generation_time
12104
tokens_prompt
77
tokens_completion
859
native_tokens_prompt
67
native_tokens_completion
790
native_tokens_completion_images
(null)
native_tokens_reasoning
0
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"
usage
0.00023058
router
(null)
provider_responses
0
endpoint_id
"722e0746-2034-4d6e-a08f-e6271849ee9e"
id
"50750656-0112-4a3a-87be-aa652cf74fdc"
is_byok
false
latency
272
model_permaslug
"deepseek/deepseek-v4-flash-20260423"
provider_name
"DeepSeek"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.1; linux; x64))"
http_referer
(null)
request_id
"req-1777124208-DHMfX27dzrpixaiJd7Za"
session_id
(null)
api_type
"completions"
id
"gen-1777124208-vYEZwbqKyVC7RJUHTGvS"
upstream_id
"50750656-0112-4a3a-87be-aa652cf74fdc"
total_cost
0.00023058
cache_discount
(null)
upstream_inference_cost
0
provider_name
"DeepSeek"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
547
0.00%
Dialogue to Total Word Ratio
Ratio: 73.95%, Deviation: 43.95%
neededClean
false
wordsTotal
549
wordsDialogue
406
0.0000%