NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1M19ZRKMA09FT7R2RS30RN8
Run Details
Model:
Hy4 Preview (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
5m 50s
Cost:
$0.066953
Provider:
Tencent
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
98.35%
Dialogue to Total Word Ratio
(details)
99.1759%
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 van smelled of regret. Across the street, a house exploded with balloons and a banner that read, “PRINCESS SPARKLES TURNS SIX!” Molly lowered her binoculars. “Remind me why we’re robbing a birthday party instead of, I don’t know, a bank.” Bill checked his reflection in a spoon. “Banks have guards, cameras, and a guy named Steve who takes his job way too seriously. Birthday parties have juice boxes, distracted parents, and a ninety percent chance nobody notices a missing princess castle. Also, look at that pile.” He pointed at a mountain of glittery gifts by the cake table. Molly sighed. “Fine. What’s the plan?” Bill unfolded a napkin blueprint. “Step one: I go in as Chuckles the Clown, a man with a tragic fear of balloons. You go in as… what’s your alias?” “Molly the Magnificent,” she said. “I have a cape.” “Great,” Bill said. “I distract the kids with balloon animals, most of which will look like depressed snakes, while you swap the real presents for empty boxes.” Molly frowned. “Empty boxes?” “Filled with cat toys,” Bill said. “Kids love disappointment. It builds character and reduces the resale value of their childhood.” “That’s evil,” Molly said. “That’s Tuesday,” Bill said. Molly studied the yard. “How do we get past the mom? She once made a grown banker cry during a bake sale.” “Which one?” Bill asked. “The one with the clipboard and the eyes of a hawk,” Molly said. Bill grinned. “Ah, Denise. I’ll challenge her to a diaper-changing contest, winner gets the last cupcake and temporary custody of a toddler named Chaos.” Molly blinked. “You don’t know how to change a diaper.” “Neither does she,” Bill said. “It’s a bluff. Meanwhile, you grab the loot, avoid the sticky doorknob, and toss it out the bathroom window.” “Bathroom window?” Molly repeated. “It’s the only one without a safety bar,” Bill said. “And it’s next to the goodie bags.” Molly nodded. “Then what?” “We escape through the bounce house,” Bill said. She paused. “We’re hiding stolen presents in a bounce house?” “No,” Bill said, “we’re making our getaway in it. I’ll slash the stakes, the whole thing deflates, and we ride it down the driveway like a runaway raft.” Molly stared. “That is the stupidest thing I’ve ever heard.” “Thank you,” Bill said. “But I’m in,” Molly said. “One problem: the cake.” Bill paled. “The cake?” “There’s a dog guarding it, and he looks like he votes,” Molly said. They leaned forward. A retriever sat beside the cake in a party hat. Bill swallowed. “Can you bribe a dog?” Molly pulled a squeaky toy from her bag. “I can negotiate.” Bill clapped her shoulder. “Then Operation Sugar Crash is a go.” Molly checked her watch. “Showtime. Try not to get glitter in the van.” “No promises.” Bill opened the door and stepped into chaos of six-year-olds. Molly whispered, “For the record, I hate clowns.” “Then it’s a good thing you’re marrying one,” Bill said, and vanished into the party.
created_at
"2026-09-03T16:23:21.628Z"
9/3/2026, 4:23:21 PM
model
"tencent/hy4-preview-20260827"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1282
moderation_latency
(null)
generation_time
349841
tokens_prompt
75
tokens_completion
21352
native_tokens_prompt
83
native_tokens_completion
26743
native_tokens_completion_images
(null)
native_tokens_reasoning
25970
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
0.066953465
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788452601-MWXdGHKclmbeZwsX1Ifx"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788452601-Y8CXHBp9BBwgupjQsKoF"
upstream_id
"e901920c-575c-4c41-89c2-fb272532bd2e"
provider_responses
0
endpoint_id
"ad2a4093-3d2a-41d9-adf1-8b499ac8445d"
id
"e901920c-575c-4c41-89c2-fb272532bd2e"
is_byok
false
latency
1282
model_permaslug
"tencent/hy4-preview-20260827"
provider_name
"Tencent"
status
200
total_cost
0.066953465
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
98.35%
Dialogue to Total Word Ratio
Ratio: 70.58%, Deviation: 0.58%
neededClean
false
wordsTotal
503
wordsDialogue
355
99.1759%