NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-04-24-01KQ05H3EQPTJK2QSZ0HVKC9SD
Run Details
Model:
Z.AI GLM 5.1
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
2m 34s
Cost:
$0.058308
Provider:
AtlasCloud
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
49.02%
Dialogue to Total Word Ratio
(details)
24.5087%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 70% 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
Bill shifted uncomfortably in the driver's seat of the getaway van, peering through the binoculars at the sprawling mansion. A massive inflatable castle dominated the lawn, vibrating with the erratic movements of sugar-fueled children. He lowered the binoculars and sighed, dreading the conversation. "We are not doing this," Bill said. "We are absolutely not doing this, Molly." "Read the file," Molly said, slapping the folder onto the dashboard of the vehicle. She popped gum into her mouth, chewing with the confidence of someone who never doubted a terrible idea. "It is all right there in black and white." "It is a birthday party, Molly. For a seven-year-old." "A seven-year-old whose parents rented a petting zoo, a magician, and that bouncy castle. You know what that means?" "That little Timmy is spoiled?" "It means the present table is stacked. I am talking high-end electronics, gift cards, cash." "We are stealing presents from children." "Redistributing surplus. They will not even notice. Kids play with the box anyway." "They will notice when they unwrap nothing!" "That is what the magician is for. Misdirection. While he pulls a rabbit, we pull a PlayStation." "How exactly do you plan on getting past the parents?" Bill asked, rubbing his temples, trying to massage away the impending headache. "Disguises. We go in as the catering staff." "We do not have catering uniforms." "I have aprons. Close enough." "Close enough is not going to cut it when Karen asks you where the vegan sliders are." "I will tell her the dog ate them. Or I will just trip her. She will not notice." "What about the petting zoo? There is a goat." "It is a goat, Bill, not a bloodhound." "Goats eat everything! What if it eats the loot?" "Why would a goat eat a Playstation?" "Why are we stealing from children?" "It is a victimless crime. They get cake, they get tired. We get the payout. Everyone wins, Bill." "Except the kid who gets no presents." "He gets character. Builds resilience." "I want us not going to jail for grand theft Lego." "We are targeting high-ticket items. The kid gets the plushies, we get the hardware. It is a compromise." "You have compromised with a seven-year-old." "I am a fair criminal. The side gate is locked, by the way. How are we getting the haul out?" "I already cased it," Molly said, tapping the blueprint with a sharp fingernail. Her eyes sparkled with the thrill of the heist, completely oblivious to the moral implications of robbing a first grader. "There is a gap in the hedge by the tool shed. We load the duffel bags, pass them through, and you drive the van around." "A gap in the hedge? You measured it?" "I measured it. I even tested it with a watermelon." "Please tell me the watermelon made it." "Mostly. Are you in or out? I am not splitting this haul with someone who lacks vision." "I lack morals, apparently." "You lack a retirement fund. Grab an apron. We move when the clown starts making balloon animals." Bill stared at the bouncing castle, watching a tiny princess tackle a smaller child in a cape. The chaos was daunting, a swirling vortex of sugar and screaming toddlers. He gripped the steering wheel. "I hate clowns." "Everyone hates clowns. That is why no one watches them. Perfect cover. Now check the rope."
created_at
"2026-04-24T16:35:22.215Z"
4/24/2026, 4:35:22 PM
model
"z-ai/glm-5.1-20260406"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1727
moderation_latency
(null)
generation_time
153977
tokens_prompt
77
tokens_completion
13666
native_tokens_prompt
66
native_tokens_completion
13231
native_tokens_completion_images
(null)
native_tokens_reasoning
12494
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"
usage
0.05830814
router
(null)
provider_responses
0
endpoint_id
"7e919c5e-6c06-4182-adaf-2910d5d4b5a8"
id
"a63a6ddd1fb3410d869037036a002856"
is_byok
false
latency
1727
model_permaslug
"z-ai/glm-5.1-20260406"
provider_name
"AtlasCloud"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.1; linux; x64))"
http_referer
(null)
request_id
"req-1777048522-D4szgq2WjFGRduh0MUo6"
session_id
(null)
api_type
"completions"
id
"gen-1777048522-7AautWjDfnvqEl7KXRcC"
upstream_id
"a63a6ddd1fb3410d869037036a002856"
total_cost
0.05830814
cache_discount
(null)
upstream_inference_cost
0
provider_name
"AtlasCloud"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
563
49.02%
Dialogue to Total Word Ratio
Ratio: 73.78%, Deviation: 3.78%
neededClean
false
wordsTotal
572
wordsDialogue
422
24.5087%