NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1M24V1E5QMCHAA2VA4W4KXF
Run Details
Model:
Hy4 Preview (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
5m 24s
Cost:
$0.065913
Provider:
Tencent
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
99.95%
Dialogue to Total Word Ratio
(details)
99.9755%
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
The getaway van smelled like frosting and regret. Molly squinted through binoculars at the suburban house, where a crooked banner read HAPPY 5TH BIRTHDAY, TIMMY! Balloons bobbed like colorful spies, and somewhere a sugar-crazed toddler screamed with joy. Molly lowered the binoculars. "All right, Bill. Operation Cakewalk is a go. Target: the present table. Objective: every gift not wrapped in Paw Patrol paper, because those are cursed." Bill, wedged behind the steering wheel in a deflated clown suit, saluted with a whoopee cushion. "Define cursed. Are we talking mild inconvenience cursed, or the dog learns your Social Security number cursed?" Molly ignored him, scanning the yard. "The dog already knows my Social Security number. Look, the perimeter is soft. One mom, one dad, one inflatable castle, and a toddler named Cody who bites ankles like a tiny shark." Bill pulled a laminated party schedule from his pocket and tapped it with a crayon. "The magician arrives at three. We’ve got eleven minutes to become him, steal the loot, and leave before the piñata confessionals begin." Molly held up a warning finger. "Forget the juice box protocol. His dad uses juice boxes as walkie-talkies. If he offers one, don’t drink it. Last year I sipped and argued with his dog for forty minutes." Bill gasped, scandalized. "You argued with the dog? Molly, that’s practically treason." Molly adjusted her neon pony onesie with dignity. "He had persuasive eyebrows, Bill. I’ll go in as Pony Princess. You’re my assistant, Kevin. Our distraction is the classic double bluff: I’ll juggle gluten-free cupcakes while you release the doves." Bill blinked twice. "Doves? We don’t have doves. We have three pigeons, a bag of croutons, and a lawsuit waiting to happen." "Fine, release the pigeons and yell ‘Look, a parrot!’ Then I’ll fake a fainting spell. While they clap, swap the table with decoy boxes filled with broccoli and regret." Bill frowned, horrified. "Why broccoli? Why not something lovable, like kittens?" "Because kids fear broccoli, Bill. They’ll assume the presents were healthy all along and never question it. It’s airtight." "Molly, we are about to rob a five-year-old who still believes in the Tooth Fairy and thinks his dog is a talking princess." Molly zipped her neon pony onesie with a dramatic flourish. "Relax. We’re only stealing the gifts kids don’t want. Socks, educational puzzles, the board game where you learn taxes. Basically childhood disillusionment." Bill nodded slowly, defeated. "So, the entire table." "Exactly. Put on your fake mustache and remember the safe word." Bill straightened his clown collar. "Is it ‘nap time’?" "No. The safe word is ‘uncle.’ If Cody bites you, you’ll need one." Bill opened the van door, paused, and looked back with wide, worried eyes. "Molly, if we get caught, will Timmy share his cake with us in prison?" Molly revved the engine and flashed a glittery grin. "Only if we tell him the broccoli was moral. Now move, Kevin. Childhood waits for no criminal." The van door slammed shut behind them.
created_at
"2026-09-03T16:38:01.525Z"
9/3/2026, 4:38:01 PM
model
"tencent/hy4-preview-20260827"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1381
moderation_latency
(null)
generation_time
324262
tokens_prompt
75
tokens_completion
20458
native_tokens_prompt
83
native_tokens_completion
26327
native_tokens_completion_images
(null)
native_tokens_reasoning
25563
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
0.065913049
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788453481-HatmJUdf6enDBoX964az"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788453481-CEKvVzNp5HY4GlDOdwzD"
upstream_id
"9ccafc6e-a023-489a-836e-5c509d4d2cc0"
provider_responses
0
endpoint_id
"ad2a4093-3d2a-41d9-adf1-8b499ac8445d"
id
"9ccafc6e-a023-489a-836e-5c509d4d2cc0"
is_byok
false
latency
1381
model_permaslug
"tencent/hy4-preview-20260827"
provider_name
"Tencent"
status
200
total_cost
0.065913049
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
99.95%
Dialogue to Total Word Ratio
Ratio: 70.10%, Deviation: 0.10%
neededClean
false
wordsTotal
505
wordsDialogue
354
99.9755%