NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1M1HJJ08RCTMM0AGEMT0Y0A
Run Details
Model:
Hy4 Preview (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
5m 45s
Cost:
$0.067984
Provider:
Tencent
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
93.05%
Dialogue to Total Word Ratio
(details)
96.5265%
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
The van’s windshield framed the house like a terrible painting: too many balloons, one sad banner, and a bounce house trembling as if it regretted every life choice. Molly sat in the passenger seat, studying the party through a pair of binoculars that had been stolen from a birdwatcher who never saw it coming. Bill was in the back, wrestling a roll of duct tape and a bag labeled “Absolutely Not a Kidnapping.” He lost the fight with the tape. It wrapped around his wrist, his elbow, and, briefly, his sense of self. Molly ignored him. The plan needed focus, and focus required her to pretend Bill was a quiet professional instead of a human disaster with sticky fingers. Outside, a child in a plastic crown sprinted past the front window, screaming about cake. Another child followed, wielding a juice box like a tiny mugger. Molly checked her watch. The gift table would be unguarded during the musical chairs massacre, which lasted approximately four minutes and produced approximately nine casualties. “Blue backpack,” Molly said. “Envelopes, gift cards, cash. The good stuff.” Bill finally freed himself and crawled forward. “What about the pony?” “The pony is a problem,” Molly admitted. “A miniature one. Timmy’s parents rented it. It’s wearing sequins.” “Sequins,” Bill repeated, horrified. “Yes. And it bites.” Bill considered this. “Can we bribe it with a cupcake?” “That’s your plan? Bribery and frosting?” “It worked on the guard at the lipstick factory.” “That security guard was a golden retriever, Bill.” Bill shrugged. The movement dislodged a whoopee cushion from his jacket. It sighed loudly. Both thieves froze. Inside the house, nobody noticed. The children were too busy destroying a piñata shaped like a cartoon lawyer. Molly returned to the binoculars. “Here’s the plan. I’m the replacement magician. You’re my assistant. We grab the gifts and leave before cake.” “What if somebody asks for a magic trick?” “You pull a juice box out of your sleeve.” “I don’t have sleeves.” “Then pull it from your pants, Bill. I don’t care. Just don’t use the whoopee cushion.” Bill nodded and adjusted his tie, which was actually a necktie printed with ducks. Molly had insisted it looked trustworthy. It did not. Nothing about Bill looked trustworthy, except perhaps his shoes, and those were stolen from a clown. A small shadow fell across the windshield. The miniature pony had escaped. It stood in the front yard, chewing the “Happy Birthday” banner and staring at the van with the cold judgement of a very small horse. Molly lowered the binoculars. “New plan.” “Thank God,” Bill said. “We still steal the presents.” Bill blinked. “Really?” “Look at him. That pony has more dignity than this entire party. And I bet he knows where the good gifts are.” The pony whinnied. Somewhere inside, a child began to cry. Molly grinned, opened the van door, and stepped into the suburban battlefield. Bill grabbed the duct tape, the bags, and his courage, which was mostly duct tape by now.
created_at
"2026-09-03T16:27:30.247Z"
9/3/2026, 4:27:30 PM
model
"tencent/hy4-preview-20260827"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2172
moderation_latency
(null)
generation_time
344880
tokens_prompt
75
tokens_completion
24242
native_tokens_prompt
83
native_tokens_completion
27155
native_tokens_completion_images
(null)
native_tokens_reasoning
26446
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
0.067983877
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788452850-4aF67bGzdPabuskyQV3D"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788452850-WiiAGEOUzVAEUKzuF9D7"
upstream_id
"b304ac9e-6be4-4e50-b3d9-02dc2c0dc534"
provider_responses
0
endpoint_id
"ad2a4093-3d2a-41d9-adf1-8b499ac8445d"
id
"b304ac9e-6be4-4e50-b3d9-02dc2c0dc534"
is_byok
false
latency
2172
model_permaslug
"tencent/hy4-preview-20260827"
provider_name
"Tencent"
status
200
total_cost
0.067983877
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
93.05%
Dialogue to Total Word Ratio
Ratio: 31.20%, Deviation: 1.20%
neededClean
false
wordsTotal
500
wordsDialogue
156
96.5265%