NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-06-24-01KVX3A6HT28D2G3C3MF0J3M1F
Run Details
Model:
Z.AI GLM 5.2 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
6m 2s
Cost:
$0.101975
Provider:
Z.AI
Evaluation summary
Score
Evaluator
Details Link
99.74%
Matches word count
(details)
99.22%
Dialogue to Total Word Ratio
(details)
99.4800%
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
Molly adjusted her black ski mask and peered through the powerful binoculars. The brightly colored bounce house inflated rhythmically in the manicured backyard. Dozens of screaming children ran across the perfectly trimmed grass, their shoes leaving muddy tracks. Bill shifted uncomfortably beside her, his broad shoulders snapping several branches of the azalea bushes. He checked his cheap digital watch and sighed, brushing a ladybug off his sleeve. The smell of cotton candy drifted over the wooden fence, making his stomach rumble loudly. Molly elbowed him in the ribs to keep him quiet. A clown with enormous red shoes walked past, honking a loud horn. Bill flinched. Molly rolled her eyes and pulled out a folded blueprint of the property. She tapped the paper with a gloved finger, tracing the route to the patio. Bill grabbed a granola bar from his pocket, unwrapped it, and took a big bite. He chewed loudly while Molly glared at him. She snatched the snack and stuffed it into her tactical vest. Bill whimpered but stayed focused on the lawn. A magician was setting up a small stage near the pool. Molly checked the perimeter, noting the locations of the security cameras. They were pointed outward, completely ignoring the sprawling hedge maze that provided perfect cover. She smiled under her mask. This was going to be easier than stealing candy from a baby, assuming the baby was not surrounded by wild and sugared first graders. Bill retrieved a backup granola bar from his pocket quietly. "Are you sure this is the target?" Bill whispered. "Trust me," Molly replied. "Timmy's parents are moguls. Those presents are worth millions." "Okay, there are clowns. I hate clowns." "Everybody hates clowns. They are a perfect distraction. We slip in and slip out." "What about the guard dog?" "It is a golden retriever puppy, Bill." "Puppies have teeth, Molly." "Bring a toy. When the magician pulls the rabbit out, the kids will rush the stage. That is our window." "The magician does not start for another hour." "Then we wait." "What if the presents are wrapped?" "I have a contact inside. The nanny gave me the manifest. Lego sets, an iPad, a Tesla. The haul is legit." "The nanny is in on it?" "She gets ten percent. We bypass the gate security. It is a simple keypad." "Do you have the code?" "One, two, three, four. These people have money, but zero imagination." "What if it is a trap?" "It is a trap set by a kid, Bill. We get hit with a foam sword." "I do not know. It feels wrong." "We are not stealing from children. We are redistributing wealth from moguls." "Until Timmy cries." "Timmy has a backup iPad. Put on your face paint." "Why do I have to wear face paint?" "Because if we get caught, we are two lost clowns looking for the circus." "I am not putting on a red nose." "Put on the nose, Bill." "Fine. But if a toddler throws up on my shoes, I quit."
created_at
"2026-06-24T15:17:24.682Z"
6/24/2026, 3:17:24 PM
model
"z-ai/glm-5.2-20260616"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3429
moderation_latency
(null)
generation_time
360492
tokens_prompt
77
tokens_completion
18576
native_tokens_prompt
73
native_tokens_completion
23153
native_tokens_completion_images
(null)
native_tokens_reasoning
22502
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
0.1019754
router
(null)
provider_responses
0
endpoint_id
"442ea97f-ad5e-40d5-b9a6-66e9e0417dce"
id
"20260624231726d274a22f635b4dc1"
is_byok
false
latency
3429
model_permaslug
"z-ai/glm-5.2-20260616"
provider_name
"Z.AI"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.16.0; linux; x64))"
http_referer
(null)
request_id
"req-1782314245-CVvG9zB49LtxHyUsiuSt"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1782314245-SO3xPshuePeLElBRwevM"
upstream_id
"20260624231726d274a22f635b4dc1"
total_cost
0.1019754
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Z.AI"
response_cache_source_id
(null)
data_region
"global"
Evaluation details
Result
Evaluator
Details
Meta Data
99.74%
Matches word count
n/a
neededClean
false
words
504
99.22%
Dialogue to Total Word Ratio
Ratio: 49.60%, Deviation: 0.40%
neededClean
false
wordsTotal
504
wordsDialogue
250
99.4800%