NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-03-25-01KMJ7WEPX7895E50FQN04TCSZ
Run Details
Model:
Qwen 3 32B
Test:
Dialogue tags
Scenario:
Write 200 words with 10% dialogue
Duration:
9.4s
Cost:
$0.000334
Provider:
Novita
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
0.0001%
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
Molly crouched in the treehouse, her night-vision goggles slipping over a neon-green pacifier. Bill, clutching a stolen walkie-talkie, nudged a inflatable dinosaur with his boot, sending it wobbling. “Phase one: disable the balloon army,” she whispered, lobbing a mini siren to the floor. It buzzed to life, emitting *Baby Shark*. Bill gaped. “This is a trap. They’ve replaced the security system with *toddler chaos*.” “The glitter isn’t a decorator’s choice,” Molly hissed, swiping a motion sensor. “It’s a particulate-based alarm. We move, we glow like a Christmas tree.” A(piñata swung overhead, trailing confetti. Bill backpedaled into a cake stand, triggering a cascade of cupcake humidity. “We’re in a dessert warzone!” “Remember the rules,” Molly said, slinging a grappling hook. “Steal the gifts, exit via zip line, no harm to—” “Molly, there’s a firewall made of Duplo blocks!” Bill yelped, as a toddler in a恐龙 onesie toddled toward him, clutching a sticky lollipop. She exhaled sharply. “Duplos are child’s play. You’re the one who thought ‘silly string’ was a *weapon*.” A kazoo band erupted from the wall. Bill flinched. “I thought we were robbing a party, not casting *Sesame Street*!” Molly’s goggles flickered. “Zip line’s in three… two…” A slipper collision sent them crashing through a piñata, bathing them in rainbow sprinkles. Below, a crowd of children gasped. “—And one!” Bill finished, wincing as a dozen tiny hands pointed. Somewhere, a kazoo played *Jailhouse Rock*.
created_at
"2026-03-25T10:15:45.634Z"
3/25/2026, 10:15:45 AM
model
"qwen/qwen3-32b-04-28"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
281
moderation_latency
(null)
generation_time
9277
tokens_prompt
93
tokens_completion
810
native_tokens_prompt
73
native_tokens_completion
725
native_tokens_completion_images
(null)
native_tokens_reasoning
381
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
0.00033355
router
(null)
provider_responses
0
endpoint_id
"d342f1c5-34b5-4a5c-b2ea-a0cdcbf72e70"
id
"44fe180a-9781-4684-93e0-b2a47dd9e92e"
is_byok
false
latency
281
model_permaslug
"qwen/qwen3-32b-04-28"
provider_name
"Novita"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.0; linux; x64))"
http_referer
(null)
api_type
"completions"
id
"gen-1774433745-Ru7AKYMIHYwGK5C2fxGK"
upstream_id
"44fe180a-9781-4684-93e0-b2a47dd9e92e"
total_cost
0.00033355
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Novita"
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
234
0.00%
Dialogue to Total Word Ratio
Ratio: 39.92%, Deviation: 29.92%
neededClean
false
wordsTotal
238
wordsDialogue
95
0.0001%