NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-04-25-01KQ2GB8BSADWG9KGR1982413E
Run Details
Model:
DeepSeek V4 Pro
Test:
Dialogue tags
Scenario:
Write 200 words with 10% dialogue
Duration:
31.3s
Cost:
$0.000285
Provider:
DeepSeek
Evaluation summary
Score
Evaluator
Details Link
95.99%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
47.9934%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 200 word scene that has 10% 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 centrepiece of the party, a piñata shaped like a woefully optimistic unicorn, listed slightly in the evening breeze, its pastel flank packed, so Molly’s fence had breathlessly informed them, not with boiled sweets but with a dozen misdirected limited-edition smartwatches. Bill surveyed the unicorn through his night-vision monocular, but his focus kept drifting to a bouncy castle shaped like an unlicensed cartoon dinosaur. “Right,” Bill whispered, “so I’ll create a diversion by ‘accidentally’ tripping the main fuse box with a frisbee.” Molly lowered her own binoculars, a frown creasing her face mask. “A frisbee, Bill?” “Aerodynamic and untraceable.” “You famously cannot throw a frisbee. You once hospitalised a vicar with a distance record attempt.” Bill sniffed, wounded. “He walked into my flight path.” “So we’re agreed, no projectile-based diversions from you,” Molly concluded, turning back to the unicorn. “We stick to the plan. I’ll approach from the sandpit, cutting the main tether with the cake knife. You will simply provide the agreed-upon childish shrieking from the hedge to cover the noise of the unicorn’s tragic deflation.” She patted his arm. “You have a very annoying shriek, Bill. It’s your best asset.”
created_at
"2026-04-25T14:23:14.694Z"
4/25/2026, 2:23:14 PM
model
"deepseek/deepseek-v4-pro-20260423"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1504
moderation_latency
(null)
generation_time
10405
tokens_prompt
77
tokens_completion
329
native_tokens_prompt
67
native_tokens_completion
294
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
0.000284925
router
(null)
provider_responses
0
endpoint_id
"d084170b-dd99-4e53-9ae1-b63881f3cfbf"
id
"f57d84f7-78b3-49fd-8fff-e4176e3dc2c5"
is_byok
false
latency
277
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-1777126994-AmqipxjcEky31v50jGHU"
session_id
(null)
api_type
"completions"
id
"gen-1777126994-vFmfYTgJLgDp2gaPOVMG"
upstream_id
"f57d84f7-78b3-49fd-8fff-e4176e3dc2c5"
total_cost
0.000284925
cache_discount
(null)
upstream_inference_cost
0
provider_name
"DeepSeek"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
95.99%
Matches word count
n/a
neededClean
false
words
192
0.00%
Dialogue to Total Word Ratio
Ratio: 52.55%, Deviation: 42.55%
neededClean
false
wordsTotal
196
wordsDialogue
103
47.9934%