NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-02-01M1H9A9RJ4N0PV8DVJFAHGQT0
Run Details
Model:
Qwen 3.7 Flash (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
1m 2s
Cost:
$0.001299
Provider:
Alibaba
Evaluation summary
Score
Evaluator
Details Link
0.49%
Matches word count
(details)
48.68%
Dialogue to Total Word Ratio
(details)
24.5853%
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
The air in the banquet hall hung thick with the cloying scent of vanilla frosting and the distinct, ammonia tang of uncontainable childhood energy. Above the ceiling tiles, within a ductwork system that felt less like an infiltration vector and more like a rusted throat, Molly pressed her body flat against the corrugated metal. Her tactical gear, usually matte black for optimal shadow blending, had been forced into a compromise involving bright yellow rubber ducks stitched onto the shoulder pads after a misunderstanding with a mascot costume rental agency two days prior. The target lay directly below: the Gift Table of Doom. It was a sprawling mountain of rectangular ambitions, wrapped in paper that screamed of last-minute panic and discount retail therapy. Beside her, Bill was vibrating. It wasn't the subtle tension of anticipation that marked a seasoned professional; it was the erratic oscillation of a man who had consumed three espresso martinis right before donning his tuxedo tails. He was hyperventilating into a silk handkerchief embroidered with the initials 'B.B.' Molly manipulated the magnetic decoder, feeding the code sequence into the electronic lock mechanism concealed behind the drywall. The device chirped a satisfied green tone, confirming that the security override was active. She prepared the extraction rope, testing the tensile strength of the fibers woven from recycled parachute material. Bill produced a slingshot loaded with marshmallow projectiles, insisting they were necessary for crowd control in the event of a hostile parent intervention. Molly confiscated the projectile weapon and replaced it with a walkie-talkie tuned to the frequency of the DJ playing Baby Shark. The vibration of the bass shook dust from the rafters, coating Molly's eyelashes in a fine gray powder. She checked her chronometer, the face cracked from a previous skirmish involving a vending machine and a rogue golf cart. The second hand swept past the twelve with cruel precision. Four minutes remained until the designated window of opportunity. Bill finally stopped flapping the handkerchief, wiped his brow with a trembling elbow, and adjusted his crooked bow tie while muttering apologies to the architecture, his breath fogging the mesh. "Is the badger mandatory?" "It is a psychological anchor. Essential." "I have the sporks ready." "Double-tape the handles immediately." "No glitter bombs?" "Glitter triggers the silent alarms." "Doggy disco route viable?" "Inflatable tunnel smells of collective shame." "Verify the middle name." "Arthur. Definitely no siblings listed." "Gaming console rumor credible?" "Chocolate egg priority. Dumpster protocol activated." "Consider charity aspect?" "Professional conduct required. Enter the vent." "Bill, stop shaking." "Martinis. Bad timing." "Check the clock." "Three forty-four." "On my mark." "Mark received." "Execute." "Badger fits in pocket?" "Pocket bulges." "Lube canister?" "Jackhammers stuck." "Diagram useful?" "Pink is aesthetic." "Kevin eating filter?" "He consumes nylon." "Rope test?" "Fibers hold firm." "Breath check?" "Fogging mesh." "Duct stable?" "Vibration nominal." "Last warning." "Stomach growls." "Ignore hunger."
created_at
"2026-09-02T14:45:37.178Z"
9/2/2026, 2:45:37 PM
model
"qwen/qwen3.7-flash-20260727"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
406
moderation_latency
(null)
generation_time
62377
tokens_prompt
75
tokens_completion
9600
native_tokens_prompt
76
native_tokens_completion
9971
native_tokens_completion_images
(null)
native_tokens_reasoning
9276
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"
service_tier
(null)
usage
0.00129851
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788360337-6sB71y1W6lo6gOQpohdA"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788360337-BjQc9qSiDBT6ORITvO7I"
upstream_id
"chatcmpl-e7da1cf3-4b1d-9a37-b73d-f3254b459a5f"
provider_responses
0
endpoint_id
"5a9a0ee6-874a-4d03-92be-e5094c0df4e7"
id
"chatcmpl-e7da1cf3-4b1d-9a37-b73d-f3254b459a5f"
is_byok
false
latency
406
model_permaslug
"qwen/qwen3.7-flash-20260727"
provider_name
"Alibaba"
status
200
total_cost
0.00129851
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Alibaba"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
0.49%
Matches word count
n/a
neededClean
false
words
473
48.68%
Dialogue to Total Word Ratio
Ratio: 26.21%, Deviation: 3.79%
neededClean
false
wordsTotal
477
wordsDialogue
125
24.5853%