NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1M0C0XV4DXQAZGJ36GSGKNT
Run Details
Model:
Hy4 Preview (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
4m 54s
Cost:
$0.057355
Provider:
Tencent
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
99.80%
Dialogue to Total Word Ratio
(details)
99.9025%
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 suburban house glowed like a sugar coma. Inside, a dozen tiny humans screamed over a cake shaped like a dinosaur wearing a party hat. Outside, Molly and Bill crouched behind a plastic flamingo. Molly adjusted her night-vision goggles, which were actually pink sunglasses with the lenses popped out. “Bill, look at all those presents. It’s like Christmas threw up in there.” Bill squinted through binoculars that had a juice-box straw taped to the side. “I see wrapping paper, ribbons, and one box that is definitely vibrating. Do you think it’s a puppy?” “It’s either a puppy or a remote-controlled car with commitment issues. Either way, it’s coming with us.” Bill pulled a crumpled blueprint from his pocket. “According to this, we enter through the side gate, disable the toddler alarm—” “The what?” “The toddler alarm. It’s a motion sensor, a scream, and a juice box rigged to trip you.” Molly nodded. “Okay, I’ll distract the parents with my award-winning impression of a responsible adult, while you sneak in dressed as a magician.” Bill held up a glittery cape. “I have a rabbit. His name is Alfonso. He’s an inside man.” “He’s a rabbit.” “Exactly. Nobody suspects a rabbit.” Molly checked her watch. “We go in five. Remember, the objective is all the presents. No detours for cake, no adopting the puppy, and no getting emotionally invested in the magician.” Bill frowned. “But Alfonso has backstory.” “Five minutes, Bill. Five minutes.” Bill rummaged through a duffel bag shaped like a smiling cabbage, knocking aside a juice box, three yo-yos, and a whoopee cushion. He produced a grappling hook made from a bent coat hanger, a slinky, and a balloon animal that looked offended by its existence. “Behold,” he said. “The infant model.” Molly blinked twice slowly. “Huh? That is a balloon giraffe with a hook.” “It’s versatile,” Bill said, “and crime is evolving.” “We’ll get arrested.” Bill grinned. “Only if the giraffe testifies.” They crept toward the gate. Somewhere inside, a cake toppled with a wet splat. Molly froze. “That’s our cue. On three. One—” “Wait,” Bill whispered. “What if the kids notice us?” Molly grinned. “We tell them we’re from the Present Redistribution Program. Kids love bureaucracy.” Bill nodded solemnly. “Brilliant. And if they cry?” “Then we bribe them with the vibrating box. Works every time.” They slipped through the gate. The back door was unlocked, because of course it was. Molly pushed it open and stepped into a hallway lined with balloons. Bill tripped over a stuffed giraffe. “Smooth,” Molly whispered. “I meant to do that. It’s a diversion. The giraffe is compromised.” “Just grab the loot.” They crept toward the living room, where a mountain of brightly wrapped presents waited like a sugary fortress. Molly pulled out a sack labeled Totally Legal Donations. Bill pulled out a second sack labeled Alfonso's Cut. “Ready?” Molly asked. Bill adjusted his cape. “Ready. For the children.” “For the children,” Molly agreed, and they charged into the party like glitter-fueled bandits.
created_at
"2026-09-03T16:06:59.777Z"
9/3/2026, 4:06:59 PM
model
"tencent/hy4-preview-20260827"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1512
moderation_latency
(null)
generation_time
293815
tokens_prompt
75
tokens_completion
19216
native_tokens_prompt
83
native_tokens_completion
22905
native_tokens_completion_images
(null)
native_tokens_reasoning
22142
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
0.057354627
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788451619-6NkpPruLcM6uDJzqYF4U"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788451619-n9zuaxl8CkHpttfFLd5u"
upstream_id
"2fdb99be-5edc-48ea-88d0-5bd72be8c7d9"
provider_responses
0
endpoint_id
"ad2a4093-3d2a-41d9-adf1-8b499ac8445d"
id
"2fdb99be-5edc-48ea-88d0-5bd72be8c7d9"
is_byok
false
latency
1511
model_permaslug
"tencent/hy4-preview-20260827"
provider_name
"Tencent"
status
200
total_cost
0.057354627
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
99.80%
Dialogue to Total Word Ratio
Ratio: 49.80%, Deviation: 0.20%
neededClean
false
wordsTotal
506
wordsDialogue
252
99.9025%