NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-04-27-01KQ77415BTG5DM5VCHGT8TPW7
Run Details
Model:
Qwen3.6 Max Preview
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
2m 44s
Cost:
$0.041103
Provider:
Alibaba
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
0.0000%
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 unrolled the blueprint across the sticky table of the diner, smoothing out a corner that smelled faintly of syrup. She tapped a red circle drawn in crayon. "Alright, Bill. Eyes on the schematic. The target is the Miller residence. Operation Sugar Rush is a go. We have a window of twenty minutes between the cake cutting and the inevitable meltdowns." Bill squinted at the paper, chewing his straw until it flattened. "Is that a bouncy castle marked as a hostile zone? Why is there a skull drawn on the slide?" "It's a tactical nightmare, Bill. High bounce factor, unpredictable civilian movement, and static electricity that can fry our comms. If you get stuck in there with a pocketful of loot, you're done. We avoid the inflatable perimeter. The slide is greased. I've seen the dad applying PAM spray." "Copy that. No bouncing. No sliding. What's the entry vector?" "The side gate. We infiltrate during the musical chairs chaos. That's when security is distracted by the crying four-year-olds and the scrambling parents. Noise cover is maximum." "Brutal. I like it. What's the disguise? Please tell me we're magicians. I've been practicing my coin vanish." "You dropped a quarter in the toilet yesterday practicing that vanish and plunged for an hour. No. We're going in as the entertainment. Specifically, the balloon artists." Bill groaned, slumping back against the vinyl booth. "Molly, my fingers are like bratwursts. I can't make a swan. Last time I tried, I made an aggressive worm that scared a toddler and caused a diplomatic incident with a grandmother." "You won't be making swans. You'll be creating a diversion. You overinflate a poodle until it pops, drawing all eyes to your sector while I sweep the gift table. The objective is high-value plastic. I'm talking the Lego Millennium Falcon, the deluxe art set, and that singing fish everyone hates but costs forty bucks." "The fish? Really? That thing haunts my dreams. Its eyes follow you." "Black market value is through the roof, Bill. Stay focused. Now, the primary threat isn't the kids. It's the Alpha Mom. Subject Karen. She wields a laminated clipboard and has eyes in the back of her head." "I've heard legends. They say she once organized a playdate with military precision and court-martialed a nanny for late juice." "Exactly. If she spots you loitering, she'll ask if you're gluten-free and then tackle you into the dunk tank. You need to keep her engaged. Ask her about the sourcing of the organic juice boxes. Compliment the charcuterie board shaped like a unicorn. It buys me ninety seconds." "Ninety seconds to grab the mountain of wrapping paper? That's tight, Molly. What about the goodie bags? Can we hit the loot station?" "Negative. Goodie bags are low yield. We're professionals, Bill. We don't steal stickers. But I need you on the piñata protocol." Bill perked up, eyes widening. "I get to smash the piñata? I've been working on my swing." "No. You are the piñata distraction. You wear the donkey costume. You dance erratically. You draw fire away from the presents." "The donkey suit? Molly, we talked about this. The tail falls off if I sneeze. And the vision slit is a millimeter wide. I'll trip over a tricycle and face-plant into the cake." "Do you want your cut or not? The fence is paying triple for unopened toys. We pull this off, and we're eating steak for a month. Real steak." Bill sighed, adjusting his collar. "Fine. I'll wear the donkey. But if there's a juice pouch leak on that suit, I'm out. And I'm keeping the singing fish. I need to conquer my fears." "Deal. Just don't try to vanish the fish. What's the exfil?" "We exit through the hedge maze. We roll to the van." "Roll? The donkey suit is velour. It attracts leaves, twigs, and shame. I'll look like a compost heap." "Better a leafy donkey than a caught thief. Just tuck and roll. Momentum is your friend." "My knees disagree. But fine. Is the van still painted like a giant cupcake?" "Camouflage, Bill. It blends with the party theme." "It has sprinkles, Molly." "Aerodynamic sprinkles. Gear up. We move at fourteen-hundred hours. And Bill? Trim your nails. You're popping balloons, not scratching the merchandise."
created_at
"2026-04-27T10:17:52.052Z"
4/27/2026, 10:17:52 AM
model
"qwen/qwen3.6-max-preview-20260420"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1220
moderation_latency
(null)
generation_time
163491
tokens_prompt
77
tokens_completion
5389
native_tokens_prompt
76
native_tokens_completion
5257
native_tokens_completion_images
(null)
native_tokens_reasoning
4240
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
0.0411034
router
(null)
provider_responses
0
endpoint_id
"ca7fc0c7-f23d-4ac5-b2e0-945be5b61787"
id
"chatcmpl-0744044e-072f-9316-bab7-cb4e920e9bf0"
is_byok
false
latency
1220
model_permaslug
"qwen/qwen3.6-max-preview-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-1777285072-QjcpDJcqz6teoy2xTrFS"
session_id
(null)
api_type
"completions"
id
"gen-1777285072-Y80vw5hhVNZkB2t2zRJD"
upstream_id
"chatcmpl-0744044e-072f-9316-bab7-cb4e920e9bf0"
total_cost
0.0411034
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
712
0.00%
Dialogue to Total Word Ratio
Ratio: 92.07%, Deviation: 22.07%
neededClean
false
wordsTotal
719
wordsDialogue
662
0.0000%