NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1KZPGHT1AZ0MBEQEC7YNTAV
Run Details
Model:
Hy4 Preview (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
5m 56s
Cost:
$0.065233
Provider:
Tencent
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
91.99%
Dialogue to Total Word Ratio
(details)
95.9948%
Overall
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
The community center smelled of frosting and impending chaos. Molly crouched behind a plastic palm, her black suit stretched over a skeleton onesie because, as she had explained, burglary required both stealth and holiday spirit. Bill crouched beside her, wheezing into a paper bag. He adjusted the bag, which had a frowny face and the words ‘Panic Later’ scrawled in marker. Through the window, forty kindergarteners swarmed a towering pile of presents like locusts. A magician in a cape was trying to saw a mother in half while her child demanded an encore. Molly lowered her binoculars. “Behold the mother lode,” she said. “Every present in that pile is basically an unguarded ATM for toddlers.” Bill dabbed sweat from his forehead. “Those aren’t ATMs, Moll. Those are biological hazards wrapped in cartoon paper. One of them just wiped a booger on a Fire Tablet.” Molly rolled her eyes and snapped open a glittery floor plan. The paper crinkled because it was actually a birthday invitation she had stolen from a mailbox. “Relax. We wait for the piñata. Kids see a stick and a donkey, they forget basic survival instincts. That’s our window. I call it the sugar-blind spot, proven science.” Bill pointed a finger at the table. A balloon arch sagged above him. “And the parents? The dad with the face paint looks like he bench-presses strollers for fun.” Molly smirked and adjusted her wrench. The wrench was a gift from her niece. “He’ll be busy livestreaming his kid’s meltdown. Besides, our distraction is better. You’re going to wear the mascot costume and start a conga line.” Bill’s mouth fell open. His eye twitched like a Morse code distress signal. “The giant cheese? Absolutely not. Last time I wore foam, a birthday girl karate-kicked me in the kneecap and called me Gouda.” Molly patted his arm with sympathy. Her gloves left glitter on his sleeve. “That’s the cost of art, Bill. You conga left, I slip through the gift tunnel. We meet at the bouncy castle with the loot.” Bill groaned and squeezed into the cheese wedge. “If I survive, I want hazard pay. And a tetanus shot. And maybe a happy ending where we adopt a puppy.” Molly checked her watch. “No puppies. Now go. When the conga hits the cake, I cut the lights and we clean out that pile like raccoons with a retirement plan.” Bill paused at the doorway, foam tail sagging. “What if the cake is vegan? I read that vegan cake is basically a brick with frosting feelings and a grudge against joy.” Molly shoved him forward. “Then use it as a stepstool. Now move, or I’m telling Timmy you ate his candy necklace.” Bill saluted, waddled toward the party, and immediately tripped over a toddler. “I meant to do that!” he yelled. The conga line formed around his fallen body. Molly shook her head, smiling. Somewhere inside, a child shrieked, “The cheese is alive!” Molly melted into the shadows, already planning the getaway.
created_at
"2026-09-03T15:55:14.887Z"
9/3/2026, 3:55:14 PM
model
"tencent/hy4-preview-20260827"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1805
moderation_latency
(null)
generation_time
355828
tokens_prompt
75
tokens_completion
22839
native_tokens_prompt
83
native_tokens_completion
26055
native_tokens_completion_images
(null)
native_tokens_reasoning
25347
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
0.065232777
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788450914-geM9Zuhg3DN4oQ1AEuUd"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788450914-rztkj8gMhbHTTvcTyUuP"
upstream_id
"01771b7d-e9b3-439f-9af9-b9e59c96cec2"
provider_responses
0
endpoint_id
"ad2a4093-3d2a-41d9-adf1-8b499ac8445d"
id
"01771b7d-e9b3-439f-9af9-b9e59c96cec2"
is_byok
false
latency
1805
model_permaslug
"tencent/hy4-preview-20260827"
provider_name
"Tencent"
status
200
total_cost
0.065232777
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
91.99%
Dialogue to Total Word Ratio
Ratio: 51.29%, Deviation: 1.29%
neededClean
false
wordsTotal
503
wordsDialogue
258
95.9948%