NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1M0E6P4N800HZWRY7NH1844
Run Details
Model:
Hy4 Preview (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
3m 51s
Cost:
$0.044770
Provider:
Tencent
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
98.78%
Dialogue to Total Word Ratio
(details)
99.3911%
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 and Bill crouched behind a fiberglass duck, the kind of lawn ornament that judged everyone. Beyond it, a suburban backyard had been transformed into a sugar-fueled nightmare of pastel balloons, streamers, and a pony wearing a party hat with the dead-eyed resignation of a creature who had seen too many cupcakes. The scent of frosting and chaos hung in the air, while a magical princess entertainer blew bubbles with the weary concentration of someone counting down the minutes. At the far end, on a table draped in glittery plastic, sat the real prize: a mountain of brightly wrapped presents. Molly lowered the night-vision goggles she had fashioned from swim goggles and green cellophane. “Remind me why we’re robbing a birthday party?” she whispered. Bill consulted a napkin labeled OPERATION SUGAR RUSH. “Because the cake is a decoy. The real treasure is that mountain of wrapped presents.” Molly raised an eyebrow. “The father’s a tech millionaire, right?” “Crypto for dogs. The dog has more money than us.” Molly sighed. “We’re stealing from kids. That’s low.” Bill shrugged. “We’re reallocating joy. Also, the robot dinosaur pays rent.” A tiny pirate wandered past, dragging a plastic sword. Bill froze. Molly pressed a finger to her lips. The pirate eyed the hedge, sneezed so hard his eyepatch fell off, then tottered toward the cake. Somewhere, a juice box rolled across the patio like a tiny, sticky tumbleweed. Molly exhaled. “Entry?” Bill pointed at the inflatable bounce castle. “Through the bounce castle. Juice-box moat, then the gift table while the magician distracts the adults.” Molly studied the chaos. A woman in a clown wig argued with the pony. A dad attempted to light candles while holding a tablet. Meanwhile, the birthday boy sat in a high chair, wearing a paper crown and covered in cake, ruling his kingdom with a plastic spoon. “Security?” Bill grinned. “A golden retriever named Sir Wags. Bribed with bacon. He’ll help.” Molly stared. “You bribed a dog, not the parents?” “Parents are chaos. Dogs are professionals.” Nearby, the piñata burst, raining candy and shrieks. Bill checked his watch. The candy cascade bought them exactly eleven seconds of adult distraction, a miracle Bill had calculated down to the last gumdrop. “Besides, the piñata already weakened structural integrity.” Molly rolled her eyes. “Code words?” “If I say ‘nap time,’ run.” Molly smirked. “What if I say ‘glitter emergency’?” Bill adjusted his stolen clown nose. “Then drop the basket and dive.” Molly hoisted an empty laundry basket labeled ‘Unexpected Gifts.’ “Last time you said that, a toddler sat on you.” “Tactical sitting. It worked.” Molly snorted. “Fine. Let’s steal the presents and vanish.” Bill adjusted his grip on the basket. “Legally questionable, financially sound.” Molly allowed a small smile. “Happy birthday to us.” They crept toward the bounce castle as the off-key singing of “Happy Birthday” began, loud enough to cover two thieves doing an undignified waddle toward a pile of loot. Behind them, the pony escaped its hat and trotted off.
created_at
"2026-09-03T16:08:11.21Z"
9/3/2026, 4:08:11 PM
model
"tencent/hy4-preview-20260827"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1899
moderation_latency
(null)
generation_time
230766
tokens_prompt
75
tokens_completion
16063
native_tokens_prompt
83
native_tokens_completion
17873
native_tokens_completion_images
(null)
native_tokens_reasoning
17124
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
0.044769595
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788451691-RlpqkLrw4SjojQgDZCx6"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788451691-D33rEYTlFf286SoAYKBD"
upstream_id
"9021ebde-9fc6-42c2-a8a2-2d4aace39e06"
provider_responses
0
endpoint_id
"ad2a4093-3d2a-41d9-adf1-8b499ac8445d"
id
"9021ebde-9fc6-42c2-a8a2-2d4aace39e06"
is_byok
false
latency
1899
model_permaslug
"tencent/hy4-preview-20260827"
provider_name
"Tencent"
status
200
total_cost
0.044769595
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
98.78%
Dialogue to Total Word Ratio
Ratio: 29.50%, Deviation: 0.50%
neededClean
false
wordsTotal
505
wordsDialogue
149
99.3911%