NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-03-01M1M123Y24KMGW72KW8DQKCMG
Run Details
Model:
Hy4 Preview (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
5m 58s
Cost:
$0.071803
Provider:
Tencent
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
99.72%
Dialogue to Total Word Ratio
(details)
99.8585%
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
Molly crouched behind a bush, binoculars pressed, and watched toddlers dismantle a gnome. Behind her, Bill wrestled with a map chewed by a teacup pig. The target was the Whitaker house, where a six-year-old’s party hit peak present-opening. The plan, scrawled on a pizza receipt, needed distraction, deception, and a collapsible wagon. Bill flattened the map on their getaway minivan’s hood, which smelled of apple juice and regret. He traced a route with a crayon, since nice pens were for forgery. Molly joined him, chewing a stolen licorice rope. “We go in as clowns,” Molly said. “Nobody suspects a clown carrying a laundry bag.” Bill frowned. “Clowns get chased by dogs and uncles.” “These are suburban dogs,” Molly replied. “They wear sweaters and fear their own reflections.” Bill uncapped the crayon with his teeth and marked the gate. They would enter the backyard, where a bounce house bobbed like a jellyfish. The piñata, a lumpy dragon, hung from the oak and guarded the gift table. Molly had studied its weaknesses. Its tail was already cracked by a bat-wielding uncle. “What about the parents?” Bill asked. “They will be too busy with cake to notice us,” Molly said. “By the time Trevor screams, we will be gone.” Bill squinted. “Trevor?” “Every party has a Trevor,” Molly said. “He is the child who will fight a piñata for fun.” “We are stealing from children?” Bill asked, aghast. “We are stealing from capitalism,” Molly corrected. “And the piñata. Mostly the piñata.” Bill considered this and tapped the receipt. The adults would be sedated by cake and the terror of a sugar stampede. At three, Molly in a false mustache would ignite a glitter bomb in a bubble machine. Children would stampede toward the sparks. Bill would sprint inside, load the presents, and vanish before anyone noticed. “Stay away from the ball pit,” Molly added. “Last time you lost a shoe and your dignity.” Bill nodded, though his eyes drifted toward the bounce house. Molly recognized that look. It was the same look he’d had before stealing a Ferris wheel. She snapped her fingers. “Focus. We are thieves, not birthday entertainment.” Molly ignored him and adjusted her clown nose, which squeaked like a dying accordion. She handed Bill a bag marked Balloons and checked her watch. The party hit the chaotic crescendo they had awaited. Somewhere inside, a child began to wail. “Remember the signal,” Molly said. “When I honk the rubber chicken twice, you release the glitter.” Bill saluted. “And if the glitter fails us?” “Then we throw the juice boxes,” Molly said. “Children will follow juice boxes anywhere.” Bill grinned. “I love this plan. What could go wrong?” Molly did not answer, because experience had taught her the universe loved that question. She shouldered her bag and headed for the gate. Bill hurried after her, wagon wheels whispering across the pavement. He wore oversized shoes and a rainbow wig but forgot pants. Molly decided not to mention it. Professionalism had its limits.
created_at
"2026-09-03T16:19:03.756Z"
9/3/2026, 4:19:03 PM
model
"tencent/hy4-preview-20260827"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1512
moderation_latency
(null)
generation_time
358289
tokens_prompt
75
tokens_completion
23315
native_tokens_prompt
83
native_tokens_completion
28682
native_tokens_completion_images
(null)
native_tokens_reasoning
27966
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
0.071802904
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788452343-COb9Ne8hj8rWypiIfb6b"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788452343-FVeuEwPgKfhcyaRTasfi"
upstream_id
"8e492fd7-9289-4a09-a15b-9a72a6e68b32"
provider_responses
0
endpoint_id
"ad2a4093-3d2a-41d9-adf1-8b499ac8445d"
id
"8e492fd7-9289-4a09-a15b-9a72a6e68b32"
is_byok
false
latency
1512
model_permaslug
"tencent/hy4-preview-20260827"
provider_name
"Tencent"
status
200
total_cost
0.071802904
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
99.72%
Dialogue to Total Word Ratio
Ratio: 29.76%, Deviation: 0.24%
neededClean
false
wordsTotal
504
wordsDialogue
150
99.8585%