NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-04-25-01KQ2FKSS9CE46V8ZXB0H9F66C
Run Details
Model:
DeepSeek V4 Pro
Test:
Dialogue tags
Scenario:
Write 200 words with 90% dialogue
Duration:
17.1s
Cost:
$0.000335
Provider:
DeepSeek
Evaluation summary
Score
Evaluator
Details Link
0.03%
Matches word count
(details)
36.79%
Dialogue to Total Word Ratio
(details)
18.4091%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 200 word scene that has 90% 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
“You know this is insane, right?” “You keep saying that, but I’m not hearing a no.” “Because it’s a yes wrapped in a clown costume and crying.” “That’s the spirit. Kids’ parties are a goldmine, Bill. Every gift is still in the box, still has the receipt basically taped to the side.” “Uh-huh. And how do we blend in? We’re two adults with the combined sun exposure of a cave salamander.” “Easy. You’re the magician. ‘Bill-o the Unbelievable.’ Your final trick is making a wagonload of presents disappear.” “My final trick will be getting arrested by a four-year-old. What’s your cover?” “Party Princess Mol-Mol. I already stole a dress. It’s got so much tulle I could fit a flatscreen under it and still look historically accurate.” “You are not shoving a Little Tikes kitchen under your skirt.” “You’re right, that’s your job, Bill-o. You saw the tip on FaceSpace, the family just leaves the gift table by the back door. It’s a smash-and-grab with cake.” “Fine. But I’m not making balloon animals.” “Of course not. Balloon animals are a gateway art. Today, a poodle. Tomorrow, you’re accidentally making a balloon shiv for a birthday boy named Kyle.” “He’s definitely a Kyle. Fine. Let’s go ruin Kyle’s fifth birthday and our criminal records.” “That’s my Mol-Mol. Now, can you fit this glitter cannon in your vest? The magician needs an entrance.”
created_at
"2026-04-25T14:10:10.475Z"
4/25/2026, 2:10:10 PM
model
"deepseek/deepseek-v4-pro-20260423"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1004
moderation_latency
(null)
generation_time
11825
tokens_prompt
77
tokens_completion
383
native_tokens_prompt
67
native_tokens_completion
351
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
0.000334515
router
(null)
provider_responses
0
endpoint_id
"d084170b-dd99-4e53-9ae1-b63881f3cfbf"
id
"15583152-eafe-4cc7-942f-9cf682a7ec68"
is_byok
false
latency
325
model_permaslug
"deepseek/deepseek-v4-pro-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-1777126210-MYNThJ9fkIHAs9MTOJKX"
session_id
(null)
api_type
"completions"
id
"gen-1777126210-EjOgWfiGdGDOpxSG2eDe"
upstream_id
"15583152-eafe-4cc7-942f-9cf682a7ec68"
total_cost
0.000334515
cache_discount
(null)
upstream_inference_cost
0
provider_name
"DeepSeek"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
0.03%
Matches word count
n/a
neededClean
false
words
230
36.79%
Dialogue to Total Word Ratio
Ratio: 100.00%, Deviation: 10.00%
neededClean
false
wordsTotal
237
wordsDialogue
237
18.4091%