NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1KYNDDQFF6QZQMZ6WY3NME1
Run Details
Model:
Hy4 Preview (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
5m 32s
Cost:
$0.064447
Provider:
Tencent
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
99.44%
Dialogue to Total Word Ratio
(details)
99.7219%
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 taped a map of the birthday house to the wall of their smelly tiny storage-unit hideout. She tapped the living room with a marker. Bill leaned over a lawn chair, eating a granola bar he had definitely stolen from a toddler. "We are not stealing from a six-year-old," Bill said. "That's not a heist. That's a tantrum with loot." Molly rolled her eyes. "We stole a kidney-shaped swimming pool last month." "That was from a billionaire's backyard. This is a suburban living room with balloon arches and a dog named Sir Barks-a-Lot. There's a difference." The dog's name alone could be a problem. It implied parents who labeled everything, including possibly the presents. Molly explained that the pile of gifts sat under a table guarded by a sleepy mom, a fog machine, and a tray of organic chicken nuggets. Bill groaned. "How do we get past security? The last time I faced a mom, she made me apologize to a juice box." "I have a plan," Molly said. "We pose as the entertainment. You're Magnificent Bill, the birthday magician. I'm your sullen assistant, Melancholy Molly." "Why sullen?" "Because happy assistants ask questions. Sullen assistants carry the laundry baskets. We do one trick, I yell that the piñata is a safety hazard, and while everyone runs outside, we load the presents into the van." Bill considered this. "Do you know how to make a piñata sound like a safety hazard?" "Not yet. But by three o'clock I'll have watched a documentary on YouTube and screamed at a sprinkler. The sprinkler deserved it, though. I truly regret nothing. What about disguises?" Bill pulled a crumpled tuxedo from a duffel bag. It still had a boutonniere shaped like a tiny apology. "I've got this. You can wear my ex-wife's neon yoga pants and a cape made from a trash bag." "Fashion-forward," Molly said. "What about the dog?" "I'll hypnotize him." "You can't hypnotize a dog, Bill." "I can't hypnotize a man either, but confidence and turkey jerky go a long way with mammals. If he barks, we bribe him. If he bites, we claim it's part of the show." Molly checked her watch. "Party starts at three. The kids will open presents at three-fifteen. We have a window of ninety seconds before sugar psychosis sets in and they turn into tiny criminals themselves." Bill stood and brushed crumbs from the tuxedo. "And if we get caught?" "Then we tell the parents we're performance artists critiquing childhood consumerism." Bill paused. "That's actually brilliant." "I know. Now put on the trash-bag cape. Crime waits for no fashion icon." Bill sighed and opened the storage unit door. "Fine. But if a toddler tries to negotiate with me, I'm surrendering." "Negotiate?" Molly said. "Kids don't negotiate. They trade. Offer him a juice box and he'll give you his sister." "Noted." Bill said. "Let's go steal a childhood." Molly grabbed laundry baskets. "And Bill? No improvising. Last time you improvised, we ended up funding a puppet revolution."
created_at
"2026-09-03T15:37:10.336Z"
9/3/2026, 3:37:10 PM
model
"tencent/hy4-preview-20260827"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1663
moderation_latency
(null)
generation_time
331894
tokens_prompt
75
tokens_completion
21360
native_tokens_prompt
83
native_tokens_completion
25741
native_tokens_completion_images
(null)
native_tokens_reasoning
25027
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
0.064447463
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788449830-aATHI8YzUZS6egrerq7Q"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788449830-qv6qenPM8pnpimf2Gpe3"
upstream_id
"75070181-1873-4e25-aa7f-3d45ee0ec5fd"
provider_responses
0
endpoint_id
"ad2a4093-3d2a-41d9-adf1-8b499ac8445d"
id
"75070181-1873-4e25-aa7f-3d45ee0ec5fd"
is_byok
false
latency
1663
model_permaslug
"tencent/hy4-preview-20260827"
provider_name
"Tencent"
status
200
total_cost
0.064447463
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
99.44%
Dialogue to Total Word Ratio
Ratio: 70.33%, Deviation: 0.33%
neededClean
false
wordsTotal
509
wordsDialogue
358
99.7219%