NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-16-01KHM5D8JXMJWTG2092DFN2VJN
Run Details
Model:
DeepSeek V3.1
Test:
Dialogue tags
Scenario:
Write 200 words with 50% dialogue
Duration:
4.4s
Cost:
$0.000499
Provider:
WandB
Model input
User/Human
You are a author of comedic heist books. Please write a 200 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 warehouse was dark, save for the single bare bulb illuminating a table covered in blueprints and pastry crumbs. Molly, peering through a jewelers loupe, delicately soldered a wire to a tiny music box mechanism. “The target is a seven-year-old’s birthday party,” she said without looking up. “Security is tight. Chuck E. Cheese level.” Bill, tightening the bolts on a small, wheeled dolly, scoffed. “A kid’s party? Molly, we’re artists. We stole the Crown Jewels of Luxembourg.” “The ‘jewels’ were paste, Bill. You were duped by a tourist trap.” She held up her creation. “This is a high-frequency emitter. It plays ‘Happy Birthday’ at a pitch only canines and irritable toddlers can hear. Causes maximum, beautiful chaos.” “So, the extraction?” “Simple. You go in as the balloon-animal guy. I’ll be the overly-affectionate aunt. When the emitter goes off, you grab the loot. I’ll handle the ‘distraction’.” She produced a large, industrial-sized whoopee cushion. Bill stared. “A whoopee cushion? That’s the masterstroke?” “In a room full of seven-year-olds?” Molly grinned. “It’s a weapon of mass distraction.”
created_at
"2026-02-16T21:23:59.216692+00:00"
2/16/2026, 9:23:59 PM
model
"deepseek/deepseek-chat-v3.1"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
231
moderation_latency
(null)
generation_time
4298
tokens_prompt
95
tokens_completion
316
native_tokens_prompt
68
native_tokens_completion
280
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
0.0004994
router
(null)
provider_responses
0
id
"cmpl-73c58f21d8d84080bea8e02fc7e9f023"
status
200
is_byok
false
latency
68
endpoint_id
"a50fbc5c-1405-4940-92ad-c455b0396643"
provider_name
"WandB"
model_permaslug
"deepseek/deepseek-chat-v3.1"
api_type
"completions"
id
"gen-1771277034-DGP68H8KBIRlgeL5315f"
upstream_id
"cmpl-73c58f21d8d84080bea8e02fc7e9f023"
total_cost
0.0004994
cache_discount
(null)
upstream_inference_cost
0
provider_name
"WandB"
Evaluation details
Result
Evaluator
Details
Meta Data
43.3783%
Matches word count
n/a
neededClean
false
words
183
4.0251%
Dialogue to Total Word Ratio
Ratio: 63.39%, Deviation: 13.39%
neededClean
false
wordsTotal
183
wordsDialogue
116
23.7017%