NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2024-09-20-01J8819N4MMX2H2BW2NQV1TV73
Run Details
Model:
Qwen 2.5 72B
Test:
Dialogue tags
Scenario:
Write 200 words with 90% dialogue
Duration:
12.4s
Cost:
$0.000152
Provider:
Hyperbolic
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
100.00%
Dialogue to Total Word Ratio
(details)
99.9994%
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
"Bill, I'm telling you, this is the perfect gig," Molly whispered, eyes gleaming with mischief. "A kid’s party loaded with unsupervised presents? It’s like a gift from the universe." Bill leaned back, crossing his arms. "And what, exactly, are we supposed to do with a bunch of toys and candy, Molly?" "Resell them, genius. There’s a market for everything, even plastic dinosaurs and lollipops." Bill shook his head. "That's a bit low, even for us. Kids are involved, remember?" "Bill, we’re not hurting anyone. We’re just relocating the goods. Think of it as a redistribution of wealth." "Redistribution, huh? What’s our cover story if someone asks why we’re in a bouncy castle?" "We’ll say we’re clowns. Professional clowns hired to entertain the kids. No one will question a guy in a fake nose." "Right, because I look great in a polka-dotted jumpsuit." "Bill, this is easy money. And think about it—no security cameras, no high-tech alarms. Just a bunch of sugared-up kids who won’t remember a thing after the cake." Bill sighed, reluctantly. "Fine. But if I get arrested for stealing Toys 'R' Us rejects, I’m blaming you." Molly grinned. "Deal. Now, let’s go find some face paint."
id
"gen-TMPXkvqQDGcw7pOythot5lvsVG1i"
total_cost
0.000152
upstream_id
"chat-375bca55b4d54a4487ddac014bb33f21"
created_at
"2024-09-20T15:37:46.424757+00:00"
9/20/2024, 3:37:46 PM
model
"qwen/qwen-2.5-72b-instruct"
app_id
182717
streamed
true
cancelled
false
provider_name
"Hyperbolic"
latency
288
moderation_latency
(null)
generation_time
11932
finish_reason
"stop"
tokens_prompt
78
tokens_completion
295
native_tokens_prompt
83
native_tokens_completion
297
native_tokens_reasoning
(null)
num_media_prompt
(null)
num_media_completion
(null)
origin
"https://ncbench.com/"
usage
0.000152
Evaluation details
Result
Evaluator
Details
Meta Data
100.00%
Matches word count
n/a
neededClean
false
words
199
100.00%
Dialogue to Total Word Ratio
Ratio: 89.60%, Deviation: 0.40%
neededClean
false
wordsTotal
202
wordsDialogue
181
99.9994%