NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-04-27-01KQ6ZFVTYNHRK4F2PJXDKWAAV
Run Details
Model:
Qwen 3.5 Plus (2026-04-20)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
2m 24s
Cost:
$0.022888
Provider:
Alibaba
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
36.92%
Dialogue to Total Word Ratio
(details)
18.4586%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 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
Molly unrolled the blueprint. The map was glitter-glued crayon depicting Mrs. Higgins' backyard, annotated with a stick figure labeled Barnaby. She tapped the paper with a grease-stained finger, adjusting her black gloves and checking the grappling hook in her purse. "Target: gift table at six. Parents distracted by magicians. Slip past bouncy castle, handle hamster, extract loot." Bill poked his coffee. He looked like a bear in a stained tank top. He stared at the donuts seeking absolution. "We broke into a bank. We stole a toupee. Why rob a six-year-old? My soul is heavy." "My soul needs the limited-edition console and the red velvet cake with excessive frosting. The cake is mandatory. Watch for jelly bean mines; the forum reports three casualties." Bill rubbed his temples, a vein throbbing as he traced the bouncy castle on the paper. He imagined his bulk destroying the inflatable joy. "Climb fence, dodge candy, fight rodent, steal tech and pastry. Exit strategy? The minivan is out?" "Inflatables. I ride you down the slide for momentum to clear the wall into the hedge. Break my fall." "Ride you? I'm two hundred pounds. We pop the slide. Giant whoopee cushion noise. Trapped in vinyl while clowns throw grapes." "Clowns are weak. Unionized and distracted by grapes. That's the window. Want the hoverboard? Neon wheels." Bill stared at the map. His shoulders slumped, then straightened. Neon wheels triggered a Pavlovian response, bypassing his conscience. He wiped crumbs from his chin, eyes narrowing with avarice. "Fine. If I bounce into the piƱata, I keep the controller. Barnaby is your problem. I hate teeth." Molly grinned, a predator sensing weakness. She rolled the map, shoved it into her pocket, and slid out of the booth, boots squeaking on the linoleum. "Deal. Black spandex. No squeaky shoes. We ride at dusk. Children weep; cake is delicious." "I hate my life," Bill muttered, though he reached for the laser pointer. The red dot illuminated his resignation as he rehearsed the slide.
created_at
"2026-04-27T08:04:31.213Z"
4/27/2026, 8:04:31 AM
model
"qwen/qwen3.5-plus-20260420"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
716
moderation_latency
(null)
generation_time
143524
tokens_prompt
77
tokens_completion
8902
native_tokens_prompt
76
native_tokens_completion
9524
native_tokens_completion_images
(null)
native_tokens_reasoning
9043
native_tokens_cached
0
num_media_prompt
(null)
num_input_audio_prompt
(null)
num_media_completion
0
num_search_results
(null)
num_fetches
(null)
web_search_engine
(null)
origin
"https://ncbench.com/"
is_byok
false
finish_reason
"stop"
native_finish_reason
"stop"
usage
0.022888
router
(null)
provider_responses
0
endpoint_id
"8b4a0fbb-6705-4f8d-a7f1-ad18ecc0edbc"
id
"chatcmpl-234ac97b-6b1b-9782-a82e-743135264100"
is_byok
false
latency
716
model_permaslug
"qwen/qwen3.5-plus-20260420"
provider_name
"Alibaba"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.1; linux; x64))"
http_referer
(null)
request_id
"req-1777277071-L9r4y8vDZo6M04fUYgCA"
session_id
(null)
api_type
"completions"
id
"gen-1777277071-MHsMqZLaFsg4kBBkR37G"
upstream_id
"chatcmpl-234ac97b-6b1b-9782-a82e-743135264100"
total_cost
0.022888
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Alibaba"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
331
36.92%
Dialogue to Total Word Ratio
Ratio: 54.46%, Deviation: 4.46%
neededClean
false
wordsTotal
336
wordsDialogue
183
18.4586%