NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-16-01KHM5ZKEYHV0NR46FMEXAP7ZY
Run Details
Model:
DeepSeek V3.2
Test:
Dialogue tags
Scenario:
Write 200 words with 90% dialogue
Duration:
8.8s
Cost:
$0.000140
Provider:
Novita
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
“It’s not a bank, Bill, it’s a kid’s party.” “Precisely! Minimal security. A clown and a donkey. I can take them.” “The donkey is for petting, Bill, not security.” “Exactly. A critical flaw. Now, we go during the magician's act. Maximum distraction.” “What’s our cover? Two uninvited adults lurking by the cake?” “We’re the party pirates,” Bill said, puffing his chest. “I’ve got eye patches.” “You’re wearing an eye patch. Singular. It’s from your Halloween ‘sexy cyclops’ costume.” “It’s a commitment to the bit. We’ll mingle, blend, then stuff the presents into the giant toy chest we bring as our ‘gift’.” “A chest? So we just walk out carrying a huge, heavy chest full of loot?” “We’ll say it’s a game. ‘Treasure Hunt.’ Kids love games.” “Parents love calling the police. How many presents are we even talking? A mountain of educational puzzles and sock bundles?” “It’s the principle, Moll. The job’s the job. High-risk, high-reward.” “The highest risk here is stepping on a rogue LEGO. The reward is a used fidget spinner.” “Fine. Forget the chest. We just eat the cake and leave.” “Now you’re talking.”
created_at
"2026-02-16T21:34:04.409045+00:00"
2/16/2026, 9:34:04 PM
model
"deepseek/deepseek-v3.2-20251201"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1230
moderation_latency
(null)
generation_time
8782
tokens_prompt
77
tokens_completion
328
native_tokens_prompt
67
native_tokens_completion
304
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)
origin
"https://ncbench.com/"
is_byok
false
finish_reason
"stop"
native_finish_reason
"stop"
usage
0.000139623
router
(null)
provider_responses
0
id
"85684e3bd082fa9d0a9fdfed9ef055aa"
status
200
is_byok
false
latency
1230
endpoint_id
"8a5fb6d3-ef71-4958-af54-269ffa3a3c6e"
provider_name
"Novita"
model_permaslug
"deepseek/deepseek-v3.2-20251201"
api_type
"completions"
id
"gen-1771277635-sE5HGr0YuiCZCdnpzOE8"
upstream_id
"85684e3bd082fa9d0a9fdfed9ef055aa"
total_cost
0.000139623
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Novita"
Evaluation details
Result
Evaluator
Details
Meta Data
86.3803%
Matches word count
n/a
neededClean
false
words
189
74.6352%
Dialogue to Total Word Ratio
Ratio: 97.35%, Deviation: 7.35%
neededClean
false
wordsTotal
189
wordsDialogue
184
80.5078%