NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-03-25-01KMJ7VPQQ6ES16CDX9TEFM5FB
Run Details
Model:
Qwen 3 32B
Test:
Dialogue tags
Scenario:
Write 200 words with 90% dialogue
Duration:
11.5s
Cost:
$0.000480
Provider:
Alibaba
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
93.56%
Dialogue to Total Word Ratio
(details)
46.7802%
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
Molly tugged a glittery banner over her face, her voice muffled. "Bill, the target’s a bounce house. You’re telling me you’re *terrified* of a inflatable." Bill adjusted his pirate hat, which kept slipping over his eyes. "It’s not the bounce house. It’s the *contents*. You think you can crack a vault guarded by ten-year-olds with candy canes? One of them might have a *real* sword. I read they’re into Pokémon now. You mess with their presents, they’ll *evolve*." Molly yanked the banner down, revealing a skeptical eyebrow. "Relax. The parents hired ‘Clown Vince’ to hide the gifts under a piñata. Classic mistake. Overcompensating for their own boring lives." "Clown Vince?" Bill shuddered. "That guy’s a *legend*. Last year he ‘accidentally’ dyed a kid’s hair blue. The mom *still* sends him birthday cards. How do we even approach this without getting tricked into blowing up a balloon animal?" "Simple. Disguise." Molly held up a My Little Pony duffel bag. "We dress as, uh… what’s the fancy word? *Costumed entertainers*." Bill peered into the bag. "Molly, this isn’t a disguise. This is a *costume bin*. Where’s the tactical gear? The *tools*? I’m not climbing into a bounce house dressed like a… a *pony*!" "Bill, it’s *purple*. Blend in. You’ll be a unicorn. Everyone loves unicorns. Even Clown Vince." "Unicorns are *girls’* costumes!" "Wrong. You’re a *villain*. The ponies need a antagonist to sell the drama. You’ll be ‘Shadow Sparkles.’" Bill groaned. "I’m going to die. I can already hear the questions. *‘ Uncle Bill, why’s your mustache shaking?’* *‘Aren’t you supposed to be nice?’*" Molly smirked. "Just stick to the script. When the kids scream, we grab the piñata. When the parents panic, we vanish into the cupcake table. Easy." "Easy? *Easy?!* What if they *recognize* me? I once donated a *kid’s library* to that school. They’ll think I’m here for *storytime*!" "Then you’ll tell them a *different* story. One where the piñata’s the hero. One where *we* get away with all the LEGOs." Bill sighed, adjusting his sequined tunic. "Legos. Of all things. You know what my mom’s going to build with these, right?" "Sure. A shrine to your face. Now *hurry*. The party starts in ten minutes. And try not to trip over your own tail."
created_at
"2026-03-25T10:15:21.084Z"
3/25/2026, 10:15:21 AM
model
"qwen/qwen3-32b-04-28"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
379
moderation_latency
(null)
generation_time
11463
tokens_prompt
93
tokens_completion
1232
native_tokens_prompt
72
native_tokens_completion
1137
native_tokens_completion_images
(null)
native_tokens_reasoning
511
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
0.00048048
router
(null)
provider_responses
0
endpoint_id
"344ccdcc-6e59-464d-89ab-8e5b92fe2adc"
id
"chatcmpl-802f7132-428e-957a-ba4c-635b79d298c4"
is_byok
false
latency
379
model_permaslug
"qwen/qwen3-32b-04-28"
provider_name
"Alibaba"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.0; linux; x64))"
http_referer
(null)
api_type
"completions"
id
"gen-1774433721-9u6yjoZ9SU0cXvvVdIEu"
upstream_id
"chatcmpl-802f7132-428e-957a-ba4c-635b79d298c4"
total_cost
0.00048048
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Alibaba"
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
376
93.56%
Dialogue to Total Word Ratio
Ratio: 84.92%, Deviation: 5.08%
neededClean
false
wordsTotal
378
wordsDialogue
321
46.7802%