NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-06-24-01KVX2Y9TEKRF2D6Q052BCMAK5
Run Details
Model:
Z.AI GLM 5.2 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
55.6s
Cost:
$0.011124
Provider:
Z.AI
Evaluation summary
Score
Evaluator
Details Link
35.00%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
17.5011%
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 balaclava, the scratchy wool rubbing against her nose. She stared at the crayon-scribbled blueprint taped to the kitchen table. "Let me get this straight. We are bypassing the Diamond Exchange, the museum, and the mob boss’s penthouse to steal... My Little Pony figurines?" Bill leaned over the table, tracing a line with a sticky finger. "It’s not just figurines, Molly. It’s a six-year-old's birthday party. That means presents. Plural. We’re talking iPads, gaming consoles, and cold hard gift cards. It’s a smash and grab of high-end electronics disguised as cardboard and wrapping paper." "And how exactly do you propose we bypass the perimeter?" she asked, squinting at a crude drawing of a trampoline. "There is no laser grid." "What about the guard dogs?" "It’s a golden retriever named Biscuit." Molly sighed heavily, rubbing her temples with gloved fingers. "Biscuit will still bark. We need a distraction. What did the recon show?" Bill grinned, pulling a brightly colored conical party hat from his duffel bag. "We go in as the entertainment. I’ve got a clown suit. You’re my assistant." "Absolutely not. I am not wearing face paint for a birthday heist." "It’s water-based! And think about it. The kids are distracted by my balloon animals, Biscuit gets a treat, and you slip into the backyard where the present table is." "And the parents?" "Drinking wine in the kitchen. It’s a Saturday afternoon in the suburbs, Molly. They’re practically sedated." Molly leaned back in her folding chair, chewing on her thumbnail. She hated suburban jobs. Too many unpredictable variables, like sugar-crashed toddlers and neighborhood watch captains. Still, Bill wasn’t wrong about the loot. Suburban parents routinely went into debt trying to impress each other with oversized gift bags. Their fence downtown would give them top dollar for untouched Legos and a brand new Kindle. "Fine. But I’m not wearing the oversized shoes. Give me the magician's cape." "Deal. But you have to saw me in half." "I might actually cut you in half if this goes sideways. Where is the exit strategy?" "The side gate. I’ll leave it unlatched during my juggling routine. You load the swag into the unmarked van, drive around the block, and pick me up at the climax of the show." "What's the climax of the show?" "I cry." Molly blinked. "You cry?" "Clowns crying makes parents uncomfortable. They’ll rush me out with a check just to make it stop. It's foolproof." Molly pulled the party hat from his hands and crushed it into her coat pocket. "If Biscuit bites me, you're footing the rabies shots. Let's move. We have a bouncy castle to breach before the cake is cut." Bill clapped his hands together, bouncing toward the door. "This is going to be the greatest heist since the Nice Bank of Geneva." "Just remember to tie your balloon animals fast. We aren't there to make friends."
created_at
"2026-06-24T15:10:54.811Z"
6/24/2026, 3:10:54 PM
model
"z-ai/glm-5.2-20260616"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2897
moderation_latency
(null)
generation_time
54884
tokens_prompt
77
tokens_completion
2826
native_tokens_prompt
73
native_tokens_completion
2505
native_tokens_completion_images
(null)
native_tokens_reasoning
1845
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
0.0111242
router
(null)
provider_responses
0
endpoint_id
"442ea97f-ad5e-40d5-b9a6-66e9e0417dce"
id
"2026062423105517d2e7c9398a4035"
is_byok
false
latency
2896
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-1782313854-7mrwAJOlW4x0IPy9ZPIE"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1782313854-6cTl2eF5amemM3zwZFT0"
upstream_id
"2026062423105517d2e7c9398a4035"
total_cost
0.0111242
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
35.00%
Matches word count
n/a
neededClean
false
words
482
0.00%
Dialogue to Total Word Ratio
Ratio: 67.62%, Deviation: 17.62%
neededClean
false
wordsTotal
488
wordsDialogue
330
17.5011%