NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-04-27-01KQ6ZEJMJVNF9P72B03JDNZ43
Run Details
Model:
Qwen 3.5 Plus (2026-04-20)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
3m 46s
Cost:
$0.036062
Provider:
Alibaba
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
33.15%
Dialogue to Total Word Ratio
(details)
16.5731%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 30% 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 tinfoil goggles while Bill wrestled a unicycle out of the trunk of the stolen bakery van. The plan was simple: infiltrate the Bouncing Barnaby bash at the Johnson residence, seize the loot, and vanish. Bill finally conquered the unicycle, clung to the roof rack, and encased himself in a polka-dot bow tie the size of a dinner plate. His shirt, reading I Paused My Games, strained against his gut. Molly smoothed her sequined jumpsuit, the material squeaking with every movement, and checked the utility belt stuffed with lockpicks disguised as magic wands. The target, the Goody Mountain, loomed near the snack table, a pyramid of gift bags guarded by Tyler, a teenager doom-scrolling so hard he missed the tricycle blocking the driveway. The afternoon sun glinted off the stolen whoopee cushions in the back. Molly raised a toilet-paper-roll periscope, her eye peering through the lens to gauge the wind and the angle of approach. Tyler's phone screen dimmed as the low-battery warning flashed, the signal they had been waiting for. The engine coughed, sputtered, and died. Molly cranked the key. Nothing but the sound of Kevin barking at a leaf. The sun beat down on the van roof, heating the interior to the temperature of a convection oven set to broil. Sweat trickled down Bill's back, pooling in his oversized floppy shoes, creating a damp squelch with every micro-movement. "Do I look like a felon?" Bill asked. "Bill, you look like a disgraced circus act fired for excessive juggling. Fix the bow tie. It's crooked like your moral compass. Tuck in the shirt. We're doing magic. Sleight of hand and grand larceny." "The haul is mostly juice and whistles. The LEGO set is vintage 1999. Kevin has predator eyes. He suspects us." "Kevin is a simpleton. He reacts only to rapid motion and pheromones of fear. Three gummy worms buy exactly twelve seconds of distraction. Use them wisely." "Decoys are four bags of cement painted gold. They are heavy, durable, and labeled Heavy. Be Careful. Glitter only if Kevin achieves sentience. We are ghosts. We steal the LEGO set. Mount the steed." "For the plastic spoons! For glory!"
created_at
"2026-04-27T08:03:49.016Z"
4/27/2026, 8:03:49 AM
model
"qwen/qwen3.5-plus-20260420"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
700
moderation_latency
(null)
generation_time
225889
tokens_prompt
77
tokens_completion
13017
native_tokens_prompt
76
native_tokens_completion
15013
native_tokens_completion_images
(null)
native_tokens_reasoning
14499
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
0.0360616
router
(null)
provider_responses
0
endpoint_id
"8b4a0fbb-6705-4f8d-a7f1-ad18ecc0edbc"
id
"chatcmpl-d63b6b51-dd67-9756-8731-503260f0207b"
is_byok
false
latency
700
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-1777277029-H1wIy8djaDJtOwWisONE"
session_id
(null)
api_type
"completions"
id
"gen-1777277029-bddQT07HDYeMYapjscqc"
upstream_id
"chatcmpl-d63b6b51-dd67-9756-8731-503260f0207b"
total_cost
0.0360616
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
361
33.15%
Dialogue to Total Word Ratio
Ratio: 34.70%, Deviation: 4.70%
neededClean
false
wordsTotal
366
wordsDialogue
127
16.5731%