NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1KY4ERZ8Y9Y423M5YSGGHN2
Run Details
Model:
Hy4 Preview (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
4m 16s
Cost:
$0.046675
Provider:
Tencent
Evaluation summary
Score
Evaluator
Details Link
43.38%
Matches word count
(details)
99.69%
Dialogue to Total Word Ratio
(details)
71.5338%
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 stale donuts and regret. Molly spread the blueprint across the steering wheel. Bill craned his neck, then frowned. Outside, a plastic flamingo watched them from the neighbor's lawn, judging their life choices. "That's not a blueprint," Bill said. "That's a birthday invitation with a crayon dinosaur on it." "It's reconnaissance," Molly said. "The dinosaur is holding a balloon. That means there will be balloons. Balloons mean people will be looking up. We go in low." Bill tapped the paper. "We're breaking into a six-year-old's party, not a casino. What's the haul?" "Presents," Molly said. "A mountain of them. Kids get ridiculous loot now. Tablets, scooters, those plastic toys that cost more than my first car." "We're stealing from children," Bill said. "Isn't there, I don't know, a line?" Molly rolled up the invitation. "There is. And it's drawn in washable marker. Look, we're not monsters. We'll only take the unopened gifts." "Ah, yes," Bill said. "The ethical thief's favorite flavor: lightly wrong." Molly grinned. "Exactly. Now, the plan. You'll dress as a clown." "No." Bill crossed his arms and glared at a cluster of balloons tied to the mailbox. "A very sad clown." "Still no." He slumped lower in the seat, as if gravity itself could cancel the job. "Bill, the entire party is supervised by adults who will instantly trust a man in a red nose. It's basically a uniform for not being questioned." Bill sighed. "What's my motive?" "You're here to perform," Molly said. "You'll do a magic show. While you fumble a rabbit, I'll be inside filling sacks." "I don't have a rabbit." Molly rummaged through a duffel bag and tossed a lone argyle sock onto his lap. "Use a sock." Bill stared. "A sock?" "Kids love socks," Molly said. "Especially if you call it Mr. Sockington and pretend it's shy. Now, distraction number two: the piñata." "We're hitting a piñata?" The paper-mache donkey on the porch seemed to nod in silent terror. "No, we're weaponizing it," Molly said. "At precisely 3:15 you'll shout that the piñata is actually a portal. Kids will riot. Parents will panic. I'll stroll out with the loot." "A portal to where?" "Doesn't matter," Molly said. "Children have no critical thinking skills before juice box hour. Now, getaway. The wagon." Bill blinked. "A wagon?" "A Radio Flyer," Molly said. "We load it with presents and roll away whistling. Nobody questions a wagon. Wagons are innocent. Wagons are nostalgia with wheels." "What if a parent asks why a clown is stealing their child's birthday?" "Then you say it's part of the show," Molly said. "Birthday robbery theater. Very trendy." Bill rubbed his temples. "This is the dumbest plan since the great vending machine Incident." A distant chorus of children began singing, slightly off-key. "And yet," Molly said, "we're doing it. Because the alternative is another month of ramen and regret." Bill opened the van door. "Fine. But if I have to wear the nose, you're explaining to the sock why it's unemployed." "Deal," Molly said. "Now put on your face paint. We rob birthday at three."
created_at
"2026-09-03T15:27:54.663Z"
9/3/2026, 3:27:54 PM
model
"tencent/hy4-preview-20260827"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1306
moderation_latency
(null)
generation_time
256261
tokens_prompt
75
tokens_completion
15792
native_tokens_prompt
83
native_tokens_completion
18635
native_tokens_completion_images
(null)
native_tokens_reasoning
17860
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
0.046675357
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788449274-jml8Suk6Y5qJuoPfjrJc"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788449274-zGpbjVPK2hR4qd51UFNZ"
upstream_id
"57675468-e06e-4351-a6f7-51b0c3e69646"
provider_responses
0
endpoint_id
"ad2a4093-3d2a-41d9-adf1-8b499ac8445d"
id
"57675468-e06e-4351-a6f7-51b0c3e69646"
is_byok
false
latency
1306
model_permaslug
"tencent/hy4-preview-20260827"
provider_name
"Tencent"
status
200
total_cost
0.046675357
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
43.38%
Matches word count
n/a
neededClean
false
words
517
99.69%
Dialogue to Total Word Ratio
Ratio: 70.25%, Deviation: 0.25%
neededClean
false
wordsTotal
521
wordsDialogue
366
71.5338%