NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-02-26-01KJBKYP9RS3GBQXY4YZSDRSHJ
Run Details
Model:
Qwen 3.5 27B
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
1m 9s
Cost:
$0.015232
Provider:
Alibaba
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
34.12%
Dialogue to Total Word Ratio
(details)
17.0580%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 70% 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 adjusted her night vision goggles. Bill leaned against the fence, chewing on a gum wrapper. "Are you sure about this?" Bill asked. "It is a six-year-old birthday." "It is not about the age," Molly whispered. "It is about the volume of loot. Look at that pile." "Most of it is plastic swords," Bill said. "And action figures that do not come with batteries." "Exactly," Molly replied. "Parents buy them. Kids open them. We take them before the unboxing. Retail value is intact." "But the guilt," Bill groaned. "What if the kid cries?" "He will not know," Molly said. "We replace them with bricks wrapped in paper." "Bricks?" Bill laughed nervously. "That is heavy. My back is not what it used to be." "Use the van," Molly instructed. "The one with the fake bakery sign." "The bakery van?" Bill raised an eyebrow. "We are stealing presents not cupcakes." "It is about misdirection," Molly explained. "Security thinks we are delivering frosting. We are delivering nothing but empty boxes." "And the parents?" Bill asked. "They are inside eating cake." "They are distracted by the clown," Molly said. "The clown is the distraction. We are the extraction team." "I still feel bad," Bill muttered. "Taking a Transformer from a child." "It is not taking," Molly corrected. "It is redistribution. We sell them online. The money goes to well us." "Right," Bill sighed. "Just us." "Focus," Molly snapped. "The fence is unlocked. I scoped it earlier." "You scoped a suburban backyard?" Bill asked. "With a drone?" "A very small drone," Molly said. "It looked like a bee." "A bee drone?" Bill shook his head. "You are crazy." "I am profitable," Molly countered. "Now grab the crowbar. We have five minutes before the cake cutting." "Cake cutting," Bill repeated. "That is when the music stops." "That is when the parents look outside," Molly said. "We need to be in and out." "What about the dog?" Bill pointed at a sleeping pug. "Bribe it," Molly said. "I have a sausage." "You brought a sausage for a heist?" Bill asked. "It is a tactical sausage," Molly said. "Now move." Bill stepped over the hedge. Molly followed. "Wait," Bill whispered. "What if the presents are digital?" "Then we steal the tablets," Molly said. "But I checked the manifest. Physical goods only." "Good," Bill said. "Less hacking." "Less hacking," Molly agreed. "More lifting." They reached the patio door. Molly picked the lock. "In we go," Bill said. "Happy stealing," Molly said.
created_at
"2026-02-26T00:00:37.181Z"
2/26/2026, 12:00:37 AM
model
"qwen/qwen3.5-27b-20260224"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
474
moderation_latency
(null)
generation_time
68518
tokens_prompt
77
tokens_completion
6104
native_tokens_prompt
76
native_tokens_completion
6337
native_tokens_completion_images
(null)
native_tokens_reasoning
5722
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
0.0152316
router
(null)
provider_responses
0
endpoint_id
"f3f88554-0c22-4d57-973b-0dec948808e8"
id
"chatcmpl-09faa005-9396-9c45-946d-34458d11d7b6"
is_byok
false
latency
474
model_permaslug
"qwen/qwen3.5-27b-20260224"
provider_name
"Alibaba"
status
200
api_type
"completions"
id
"gen-1772064037-Cm2pIRHN5YrXY6bcnOSy"
upstream_id
"chatcmpl-09faa005-9396-9c45-946d-34458d11d7b6"
total_cost
0.0152316
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
412
34.12%
Dialogue to Total Word Ratio
Ratio: 74.64%, Deviation: 4.64%
neededClean
false
wordsTotal
414
wordsDialogue
309
17.0580%