NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1M1BKXPZ7M4Q3YCXBFD4GXT
Run Details
Model:
Hy4 Preview (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
3m 37s
Cost:
$0.041508
Provider:
Tencent
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
95.37%
Dialogue to Total Word Ratio
(details)
97.6830%
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 spread the blueprint across the van's floor and jabbed a manicured fingernail at the backyard. "According to the invitation, the piƱata is scheduled for three o'clock. That gives us a forty-five minute window while every adult is hypnotized by a man in a sweaty frog costume." Bill peered over her shoulder, crumbs from a stolen donut dotting his mustache. "A frog? Please. Last year it was a princess who cried because the dragon had union issues. I say we go in as caterers. Nobody questions people carrying trays." She rolled up the blueprint and swatted his knee. "You'd last two minutes. You're terrified of juice boxes. Remember the Henderson job? You surrendered to a Golden Retriever." "That dog had a badge." "It had a bandana." He adjusted his night-vision goggles, which were currently sitting on his forehead like a neon headband. "Fine. New plan. We wait until the cake arrives. I hide inside the giant inflatable castle, you cut the power, and we roll the presents out through the bounce house's emergency exit." Molly narrowed her eyes. "There is no emergency exit on a bounce house." "There is now." Bill produced a box cutter from his pocket and wiggled his eyebrows. "I watched a tutorial." She groaned. "We are not deflating a children's amusement ride. We need something subtle. Elegant. Something that says 'professional burglars' and not 'two raccoons in a trench coat.'" "Raccoons get a bad rap." He pulled a crumpled party schedule from his jacket. "Look, the clown takes a break at four. His name is Chuckles, but his real name is Greg, and Greg has a severe fear of balloons. I can use that." Molly snatched the paper. "You want to blackmail a clown?" "I want to borrow his identity. Greg goes in, Greg comes out, and Greg takes the loot. The kids will be so confused they'll think it's part of the show." She tapped a polished nail against her teeth. "We'd need a van. A big one. Those presents are piled like a small fortress. Little Ethan is either destined for Harvard or a life of crime." Bill grinned. "Both, probably. Eh, we leave him one thing. A slightly chewed teething ring." "Very generous." "Hey, we're not monsters." Molly sighed and opened the van door. The sound of squealing children and an off-key rendition of 'Baby Shark' drifted through the air. "Fine. But if I have to wear a red nose, you're wearing the rainbow wig." "Deal," Bill said, already pulling a honking horn from his pocket. "Try to contain your excitement." "I'm vibrating with joy," she deadpanned. "Just promise me one thing." "Anything." She stepped onto the pavement and straightened her jacket. "When Greg the Clown inevitably screams, you're the one who explains it to the parents. I'm not doing an encore." Bill slammed the van door. "Noted. Now let's go steal Christmas in July." Molly paused, shielding her eyes slowly. "If this goes sideways, I'm blaming the frog costume. And the clown."
created_at
"2026-09-03T16:24:15.037Z"
9/3/2026, 4:24:15 PM
model
"tencent/hy4-preview-20260827"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1235
moderation_latency
(null)
generation_time
217260
tokens_prompt
75
tokens_completion
13972
native_tokens_prompt
83
native_tokens_completion
16569
native_tokens_completion_images
(null)
native_tokens_reasoning
15864
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
0.041508291
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788452655-vnKC7wAuz4ofoZQRRG9E"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788452655-pRBIVn3NoAflESrVMTn1"
upstream_id
"a5646011-cc6c-43d3-8fbc-3df8a6a79d3d"
provider_responses
0
endpoint_id
"ad2a4093-3d2a-41d9-adf1-8b499ac8445d"
id
"a5646011-cc6c-43d3-8fbc-3df8a6a79d3d"
is_byok
false
latency
1235
model_permaslug
"tencent/hy4-preview-20260827"
provider_name
"Tencent"
status
200
total_cost
0.041508291
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Tencent"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
100.00%
Matches word count
n/a
neededClean
false
words
500
95.37%
Dialogue to Total Word Ratio
Ratio: 70.97%, Deviation: 0.97%
neededClean
false
wordsTotal
503
wordsDialogue
357
97.6830%