NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-16-01KHK33QE7HW53KV2BVVRBGN2G
Run Details
Model:
Qwen 3.5 397B A17B
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
3m 27s
Cost:
$0.028554
Provider:
Alibaba
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 crouched behind the overgrown hydrangeas, peeking through the dense green gaps in the suburban hedge. The backyard buzzed with the chaotic, shrieking energy of sugar-fueled six-year-olds running across the lawn. Bill adjusted his tactical vest, which looked completely ridiculous over his bright yellow striped polo shirt. He sighed loudly, checking his digital watch for the third time in a single minute. The air smelled of hot dogs and desperation. Are you absolutely sure about this specific target? Bill whispered urgently, wiping cold sweat from his furrowed forehead. It feels morally wrong stealing from toddlers, Molly. Molly shushed him violently, gesturing sharply toward the white folding table covered in wrapping paper. It is piled high with unplugged toys and thick cash envelopes, Bill. Do you actually know what the black market value of unopened Lego sets is currently? It is truly astronomical. Bill shook his head slowly, looking at the inflatable bouncing castle wobbling in the wind. We are professional criminals. We hit banks. We hit armored cars. We do not hit birthday parties. You hit the armored car last week and got caught because you stopped for a vanilla latte, Molly reminded him sharply. This is low risk. The parents are distracted by the magic show. The security is literally a clown named Sparkles. Sparkles looks highly suspicious, Bill muttered, narrowing his eyes at the entertainer juggling colorful pins. He has very alert eyes for a clown. Sparkles wore oversized red shoes that squeaked with every step he took on the pavement. Molly pulled out a crumpled blueprint drawn on a greasy diner napkin. Here is the extraction route. We go in during the cake song. Everyone sings, nobody watches the presents. We grab the bags, slide under the fence, and vanish. What about the piñata? Bill asked, pointing at the dangling papier-mâché donkey. That seems like a potential trap. It is filled with candy, not explosives, Bill. Focus on the loot. Molly zipped up her heavy duffel bag. We need this score. Rent is due tomorrow and you spent our emergency fund on that latte. I needed the caffeine to plan the bank job properly, Bill defended himself weakly. Whatever. Ready to move? Molly asked, checking her watch again. I still think we should rob a museum instead, Bill said, standing up slowly from the grass. Too much laser security. Too much walking. Molly grabbed his arm tightly. Think of the Legos, Bill. Think of the cold hard cash inside the Hallmark envelopes. Fine. But if Sparkles catches us, I am telling him it was your idea, Bill stated firmly. Molly smiled, pulling her black ski mask over her face. Just follow my lead. We hit on the chorus of Happy Birthday. Bill pulled his own mask down over his nose. This is the lowest point of my criminal career. You will say differently when you are playing with a remote control helicopter, Molly promised. They crept toward the wooden gate. The singing began off-key and loud. The wooden gate creaked ominously as Molly pushed it open just a crack. Molly signaled go. Bill hesitated, then followed. Bill held his breath until they were fully inside the yard. They slipped into the chaos unnoticed.
created_at
"2026-02-16T11:27:57.158504+00:00"
2/16/2026, 11:27:57 AM
model
"qwen/qwen3.5-397b-a17b-20260216"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
867
moderation_latency
(null)
generation_time
206817
tokens_prompt
77
tokens_completion
8911
native_tokens_prompt
76
native_tokens_completion
7919
native_tokens_completion_images
(null)
native_tokens_reasoning
7218
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
0.028554
router
(null)
provider_responses
0
id
"chatcmpl-76967552-7542-9fe2-b35d-a31fa114ff02"
status
200
is_byok
false
latency
867
endpoint_id
"180484a2-e24b-495a-ba41-b34b680d5a9f"
provider_name
"Alibaba"
model_permaslug
"qwen/qwen3.5-397b-a17b-20260216"
api_type
"completions"
id
"gen-1771241070-NxRs0N48tf1Kw9Bjn2zj"
upstream_id
"chatcmpl-76967552-7542-9fe2-b35d-a31fa114ff02"
total_cost
0.028554
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Alibaba"
Evaluation details
Result
Evaluator
Details
Meta Data
0.0000%
Matches word count
n/a
neededClean
false
words
541
0.0000%
Dialogue to Total Word Ratio
Ratio: 0.00%, Deviation: 50.00%
neededClean
false
wordsTotal
541
wordsDialogue
0
0.0000%