NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-06-24-01KVX30HSXTP0B2FF1AX04CBJ9
Run Details
Model:
Z.AI GLM 5.2 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
1m 44s
Cost:
$0.028337
Provider:
Z.AI
Evaluation summary
Score
Evaluator
Details Link
99.38%
Matches word count
(details)
80.14%
Dialogue to Total Word Ratio
(details)
89.7589%
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 dashboard clock flickered from 1:59 to 2:00 PM. Bill adjusted his binoculars, the cheap plastic strap digging into his thick neck. Through the stolen minivan’s smudged windshield, the backyard of 42 Maple Drive was a chaotic explosion of primary colors. A bouncy castle sagged depressingly in the corner, and a sad clown with questionable makeup was twisting balloons into unidentifiable shapes. "I'm telling you, Molly, this is the easiest score of our careers," Bill said. "Two dozen seven-year-olds. Zero security. One giant pile of greed." "They are presents, Bill. We're stealing from children," Molly replied, checking the magazine of her tranquilizer gun. "I thought we had standards. We robbed a casino last month." "And the casino had cameras, guards, and a boss who wanted us dead. Look over there. The parents are doing shots by the hot tub. The loot is just sitting on that picnic table, completely unguarded. Lego sets, iPads, rare Pokémon cards. We're looking at five grand minimum." Molly sighed loudly, tucking the dart gun into her tactical thigh holster. She pulled a crumpled blueprint of the backyard from the glove compartment. She traced a careful path with her finger, deliberately avoiding the automated sprinkler system and the dubious petting zoo stationed near the wooden fence. "Fine. But we stick to the plan. I go in as the backup caterer. You cause a distraction by the ice cream truck. When the kids swarm you, I bag the goods and we vanish." "What kind of distraction?" "I don't know, Bill. Trip into the bouncy castle. Start a balloon animal riot. Just keep them away from the loot." Bill rubbed his hands together, a glint of larcenous joy shining in his eyes. He reached into the cluttered backseat and pulled on a painfully bright yellow safety vest. Molly stared at him with a mixture of pity and sheer disbelief. "What is that?" "My catering disguise. I borrowed it from my cousin Gary. He works construction, but it says 'Staff' on the back. Sort of." "It says 'Stff', Bill. And it's covered in drywall dust. You look exactly like a radioactive traffic cone." "Parents at suburban birthday parties are stressed, Molly. They aren't checking my credentials. They just want someone to refill the chip bowl. Trust me on this." Molly rubbed her temples, feeling a terrible headache forming. She looked at the picnic table again, groaning inwardly at the sight. The pile of brightly wrapped boxes was genuinely tempting. A brand new PlayStation box sat right on top. She took a deep breath and grabbed the empty duffel bag from the floorboards. "If a six-year-old bites me, I'm quitting this crew." "Noted. Now, let's go make some children cry." Bill threw open the sliding van door and marched confidently toward the yard. Molly followed closely behind, shaking her head in dismay as the clown accidentally popped a balloon sword, sending a sudden wave of shrieking children running in every direction across the manicured grass.
created_at
"2026-06-24T15:12:08.522Z"
6/24/2026, 3:12:08 PM
model
"z-ai/glm-5.2-20260616"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3761
moderation_latency
(null)
generation_time
103100
tokens_prompt
77
tokens_completion
6538
native_tokens_prompt
73
native_tokens_completion
6417
native_tokens_completion_images
(null)
native_tokens_reasoning
5770
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
0.028337
router
(null)
provider_responses
0
endpoint_id
"442ea97f-ad5e-40d5-b9a6-66e9e0417dce"
id
"2026062423120892f24b580aee4be3"
is_byok
false
latency
3761
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-1782313928-L3t0N2VLBJ6fwFPmRMQz"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1782313928-LaYZv4Zo4hMX2Y546Xye"
upstream_id
"2026062423120892f24b580aee4be3"
total_cost
0.028337
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
99.38%
Matches word count
n/a
neededClean
false
words
495
80.14%
Dialogue to Total Word Ratio
Ratio: 47.90%, Deviation: 2.10%
neededClean
false
wordsTotal
499
wordsDialogue
239
89.7589%