NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-02-01M1GS2ZSDKR3VKKBRWRCYTGZ3
Run Details
Model:
Z.AI GLM 5.3 (Reasoning, Max)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
5m 46s
Cost:
$0.109246
Provider:
Z.AI
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
30.82%
Dialogue to Total Word Ratio
(details)
65.4120%
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 blueprint was spread across the kitchen table, pinned at the corners by a bag of frozen peas, a jar of pickles, and one profoundly uninterested cat. Molly stood over it wearing her reading glasses, which she reserved exclusively for planning crimes. Bill sat across from her, still in his coat, wondering when his life had come to this. "This is Pinecrest Community Hall," Molly said, tapping the paper. "One fire exit, eleven tables, and a bouncy castle in the northeast corner — our primary obstacle." "It's a birthday party, Molly. For a seven-year-old." "Her name is Kayla, she's turning eight, and she's old enough to remember faces. That makes her a witness." Molly uncapped a marker with her teeth. "The presents go on the gift table by the window. Parents are busiest during cake, two fifteen to two forty." Bill stared at the rectangle. "Twenty-five minutes?" "Generous. The Vienna job gave us eleven." "The Vienna job was a bank." "And this is harder. Children notice everything; adults notice nothing. Walk through that hall in a clown nose and every parent assumes you're entertainment." Bill rubbed his eyes, and it occurred to him — slowly, with mounting dread — exactly where this was going. There was a wig on the counter behind Molly that he hadn't noticed until now. It was enormous. It was fuchsia. It radiated a theatrical menace he couldn't articulate. "Here's my problem," he said carefully. "You want us to go in as clowns." "Magenta and Bojo," said Molly, as if that settled the matter. "Why am I Magenta?" "Because the wig suits your face." Bill opened his mouth, closed it, and glanced at the wig, which seemed to glance back. Then he pointed at a small circle on the blueprint. "What's that?" "The piñata. At two thirty, every child in that hall swarms it for candy. Sheer chaos." "Good for us?" "Terrible for us. You never engage a mob of sugar-crazed eight-year-olds, Bill. You retreat, and you flank." "Flank a piñata." "Flank the cake table." She drew a quick, confident arrow — the arrow of a woman who had flanked many things. "You load the loot bag. I keep lookout and make balloon animals. A clown who isn't making balloon animals is a suspicious clown." Bill was quiet for a long moment. He looked at the blueprint, at the wig, at the cat, which appeared to be judging them both. "Have you considered that we're stealing presents from children?" Molly set down her pen. The professional facade flickered into something almost ashamed. "Extensively," she said. "Anything handmade goes back. Anything from a grandmother goes back. We're not monsters, Bill." "That's genuinely worse somehow." "And we leave a replacement gift. Something educational, but still fun." "A consolation prize." "A puzzle cube. It builds spatial reasoning." She rolled up the blueprint with the brisk authority of a general. "Go home and sleep. Big day tomorrow. And Bill?" "Yeah?" "Practice your balloon animals. Last time, your swan looked like a medical emergency."
created_at
"2026-09-02T10:02:00.373Z"
9/2/2026, 10:02:00 AM
model
"z-ai/glm-5.3-20260816"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
3921
moderation_latency
(null)
generation_time
345783
tokens_prompt
75
tokens_completion
21441
native_tokens_prompt
73
native_tokens_completion
24822
native_tokens_completion_images
(null)
native_tokens_reasoning
24150
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
0.10924604
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788343320-vvLYQadMRZOCnC7WucSc"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788343320-mbgPjU0GohQWhjK3NB7Z"
upstream_id
"20260902180200f4c559fad6eb44fd"
provider_responses
0
endpoint_id
"db975786-e7f6-4838-9e56-3e0b77f1bdc2"
id
"20260902180200f4c559fad6eb44fd"
is_byok
false
latency
3921
model_permaslug
"z-ai/glm-5.3-20260816"
provider_name
"Z.AI"
status
200
total_cost
0.10924604
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
100.00%
Matches word count
n/a
neededClean
false
words
499
30.82%
Dialogue to Total Word Ratio
Ratio: 54.85%, Deviation: 4.85%
neededClean
false
wordsTotal
505
wordsDialogue
277
65.4120%