NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-02-01M1GV37PDD33HY292AY3VP3JC
Run Details
Model:
Z.AI GLM 5.3 (Reasoning, Max)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
5m 33s
Cost:
$0.085851
Provider:
Z.AI
Evaluation summary
Score
Evaluator
Details Link
93.65%
Matches word count
(details)
97.73%
Dialogue to Total Word Ratio
(details)
95.6921%
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 van smelled of cold pizza and ambition. Molly spread a napkin across the dashboard, pinning its corners with a lug wrench and a stolen juice box. "Here's the layout," she said, tapping a ketchup smudge. "Bounce house here. Snack table here. Presents there." Bill leaned in. He was built for vault doors and wall-scaling, and had once held a laser-grid pose for forty minutes without blinking. But he squinted at the napkin like it had insulted his mother. "That's a napkin." "It's a napkin with intel. Party starts at two, cake at three, presents at three-thirty. That's our window." Bill shifted. In fifteen years of thieving, he'd smuggled diamonds in a saxophone and outrun Interpol on a moped, but nothing had prepared him for a suburban cul-de-sac. Pastel balloons bobbed from a mailbox shaped like a lighthouse. "There'll be kids there, Molly. A dozen of them." "Kids are just small, distracted security guards." "They scream in pitches that summon adults. It's an alarm system made of lungs." Molly held up a flyer she'd printed at the library that morning. ENTERTAINMENT NEEDED, it said, in Comic Sans. Forgery was among her many talents. "Which is why the birthday mom hired us. Her magician canceled. Tragic. Mysterious." "What did you do to the magician?" "He's fine. He's tied up. I dazzle them with magic for twenty minutes, you load the loot." "What do you know about magic?" "I can pull a rabbit out of a hat." "Do you have a rabbit?" "I have a raccoon with very reasonable energy." Bill stared at her for a long moment. Behind them, in the cargo area, something rustled inside a top hat and hissed. Neither of them acknowledged it. "And the clown? Every party has a clown. Clowns can smell a heist." "Then we handle the clown first. Professional courtesy. I offer him a balloon giraffe for his silence." "You're going to bribe a clown." "Clowns are the most corruptible people in entertainment." Bill rubbed his temples, the way he did whenever a plan contained the word "bounce," and looked at the yard, where a unicorn piñata dangled from a tree like it knew something. "What about the piñata? It's basically a tripwire full of candy." "So we wait for the piñata. When it breaks, the kids go feral. That's when we move." "And the birthday girl? She'll cry her little face off, and I'll hear it for the rest of my life. Like tinnitus." Molly paused, then hauled a box from under the seat, wrapped in glitter paper with an enormous bow. For a hardened criminal, she kept a suspicious number of soft spots. "Insurance. We leave one present behind. A hand-carved rocking horse the size of a dishwasher. She'll forget the others existed." "That's... actually kind of beautiful." "I'm a professional. Now put on your bow tie. We're on in twenty." Bill sighed the sigh of a man about to be outsmarted by a six-year-old, and reached for his bow tie. Somewhere in the cargo area, the raccoon hissed its approval.
created_at
"2026-09-02T10:37:05.624Z"
9/2/2026, 10:37:05 AM
model
"z-ai/glm-5.3-20260816"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3330
moderation_latency
(null)
generation_time
333106
tokens_prompt
75
tokens_completion
15724
native_tokens_prompt
73
native_tokens_completion
19505
native_tokens_completion_images
(null)
native_tokens_reasoning
18805
native_tokens_cached
64
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
0.08585124
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788345425-9xyXDCfzUG6GXA6sLEsc"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788345425-NeE1IOaAbV2GvoXylMFn"
upstream_id
"202609021837052c2923236f0549a7"
provider_responses
0
endpoint_id
"db975786-e7f6-4838-9e56-3e0b77f1bdc2"
id
"202609021837052c2923236f0549a7"
is_byok
false
latency
3330
model_permaslug
"z-ai/glm-5.3-20260816"
provider_name
"Z.AI"
status
200
total_cost
0.08585124
cache_discount
0.00007296
upstream_inference_cost
0
provider_name
"Z.AI"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
93.65%
Matches word count
n/a
neededClean
false
words
509
97.73%
Dialogue to Total Word Ratio
Ratio: 50.68%, Deviation: 0.68%
neededClean
false
wordsTotal
517
wordsDialogue
262
95.6921%