NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-02-01M1GTAKEEJAWJP8KMMB2MGVDA
Run Details
Model:
Z.AI GLM 5.3 (Reasoning, Max)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
1m 48s
Cost:
$0.016516
Provider:
Z.AI
Evaluation summary
Score
Evaluator
Details Link
43.38%
Matches word count
(details)
6.01%
Dialogue to Total Word Ratio
(details)
24.6919%
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
Molly crouched behind a bouncy castle at the edge of Maple Park, binoculars pressed to her face despite the party being only twelve feet away. "Target acquired. We've got a clown, twelve hostiles, and a piñata of unknown origin." Bill slid down beside her, wheezing from the jog across the parking lot. "Twelve hostiles? Molly, they're children." "Children with sticky hands and excellent hearing. Never underestimate a seven-year-old, Bill. They're like tiny detectives with no concept of mercy." "Fine. What's the layout?" "Present table at three o'clock, guarded by a woman I can only describe as aggressively maternal. Pony rides in the back. The cake comes out at three, which gives us a nine-minute window." "Nine minutes?" "When the cake comes out, everyone looks at the cake. That's not opinion, Bill, that's science." Bill peeked over the inflatable barrier and immediately ducked back down. "Why are we doing this again? Last week you said we should hit the jewelry museum." "The jewelry museum had lasers, Bill. Lasers. This place has a bubble machine and a man dressed as a wizard who I'm fairly sure is high." "So what's the plan?" "I go in as entertainment staff. You're my assistant." "Assistant? I've done twenty heists." "And how many have gone well?" Bill was quiet for a long moment. "The zoo one was mostly fine." "An ostrich stole your pants, Bill." "The ostrich thing was handled." Molly pulled a rainbow wig from her duffel bag and held it out like a crown. "Put this on." "I'm not wearing the wig." "Assistant clowns wear wigs, Bill. It's protocol." "That's not protocol, that's just a wig." "You want the presents or not? There's a pony here. Do you know what ponies cost?" "You want to steal presents from children to fund a pony?" "I want to steal presents because it's easy, and the pony is a stretch goal. Focus, Bill." He snatched the wig. "Fine. But if the clown recognizes me—" "You did one kids' party in 2019. He's not going to recognize you." "He owes me money." Molly paused, binoculars lowering slowly. "The clown owes you money?" "Forty bucks and an apology. Different story. Point is, we've got history." "Then you stay on present extraction and I'll handle crowd control. When the cake comes out, you load the table into the wagon, we exit past the bubble machine, and we're gone before the wizard finishes his third trick." "The wizard does tricks?" "He's been trying to pull a rabbit from a hat for forty minutes, Bill. The rabbit is winning. It's very sad. Nobody will be watching us." Bill adjusted the wig in the reflection of her binocular lens. "What if a kid sees me?" "Then you hand them a balloon. Kids are easily bribed. It's their greatest weakness." "That feels wrong." "It's a balloon, not a felony. Now get the wagon. And Bill?" "Yeah?" "Duck. Bubble incoming." They both hit the grass as a wall of iridescent bubbles drifted overhead, one popping against the wizard's pointy hat in the distance. "That's our cover," Molly whispered. "That's beautiful. Move out."
created_at
"2026-09-02T10:23:38.453Z"
9/2/2026, 10:23:38 AM
model
"z-ai/glm-5.3-20260816"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
4029
moderation_latency
(null)
generation_time
107940
tokens_prompt
75
tokens_completion
3593
native_tokens_prompt
73
native_tokens_completion
3747
native_tokens_completion_images
(null)
native_tokens_reasoning
3033
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
0.01651604
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788344618-5WNtsf0bmTe0pW5mdcPl"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788344618-v3kK0Z24PwioCEJP8qwc"
upstream_id
"20260902182338f0f88ad8cbe74288"
provider_responses
0
endpoint_id
"db975786-e7f6-4838-9e56-3e0b77f1bdc2"
id
"20260902182338f0f88ad8cbe74288"
is_byok
false
latency
4029
model_permaslug
"z-ai/glm-5.3-20260816"
provider_name
"Z.AI"
status
200
total_cost
0.01651604
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
43.38%
Matches word count
n/a
neededClean
false
words
517
6.01%
Dialogue to Total Word Ratio
Ratio: 77.50%, Deviation: 7.50%
neededClean
false
wordsTotal
520
wordsDialogue
403
24.6919%