NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-02-01M1H5BWSZSG7AS0QW919WX337
Run Details
Model:
Z.AI GLM 5.3 Flash (Reasoning, Low)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
28.9s
Cost:
$0.000182
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 30% 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 the binoculars and squinted through the hedge. Through the kitchen window of the Hendersons' suburban home, she could see a balloon arch, a cake shaped like a dinosaur, and a table already sagging under wrapped presents. "Bill, I'm having second thoughts about this," she said. Bill crouched beside her, wearing a party hat he insisted was part of his disguise. "It's reconnaissance, Mol. This is what professionals do." "We're professionals robbing a seven-year-old's birthday party." "Which happens to have the biggest gift haul in the neighborhood. Little Timmy Henderson is an only grandchild on both sides. Do the math." Molly lowered the binoculars. "The math is that we're going to prison, and the headline will say we stole a Lego set from a child." "Twelve Lego sets. The kid's parents are overcompensating for the divorce." Bill pulled out his notebook, which was mostly doodles. "Here's the plan. You go in as the clown." "I am not being the clown." "The clown already cancelled. I called. I was very convincing." "You called and cancelled a child's clown so you could be the clown?" "So *we* could run the play, yes." He flipped to a page with what appeared to be a crude drawing of a balloon dog. "Phase one: entertainment. Phase two: I juggle, you bag the presents. Phase three: exit through the bouncy castle. It's a natural getaway vehicle." "A bouncy castle is not a vehicle, Bill." "It bounces, doesn't it? Mol, have you ever actually robbed anything before, or just talked about it a lot?" "I've done eleven years of legitimate security consulting, which is why I know every step of this plan is insane." "Insane like a fox." Bill snapped his notebook shut. "Look, we do this one job, we split the haul, and I finally pay off Derek." "Who is Derek?" "My fence. Or he might be my landlord. Honestly the relationship is complicated." A stream of children ran past the hedge, and both thieves froze in their crouches. Molly watched a tiny girl in a princess dress march by with the confidence of a general. "Okay, new consideration," Molly whispered. "There are thirty children in there. Thirty witnesses under the age of ten, and children are terrifyingly honest." "That's why the clown disguise. Nobody suspects a clown." "Everyone suspects a clown. That's the whole cultural problem with clowns." Bill waved this off. "Fine. You be the magician's assistant instead. I got a top hat and everything." "Where is the magician?" Bill was quiet for a moment. "Potentially also cancelled." "Bill." "He had a conflicting engagement, Molly, it's not like I threatened him." "You threatened him." "It was a light threat. Very casual. Friend-to-friend." Molly put the binoculars in her bag and stood up, brushing leaves off her knees. "I'm doing this because I owe you for the boat incident, and after this we're even. Completely even. You never speak of the boat again." Bill grinned and straightened his party hat. "Deal. Now let's go steal some presents. I'll start the getaway vehicle." "It's a bouncy castle, Bill." "It's a *fast* bouncy castle."
created_at
"2026-09-02T13:36:35.145Z"
9/2/2026, 1:36:35 PM
model
"z-ai/glm-5.3-flash-20260826"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1847
moderation_latency
(null)
generation_time
28835
tokens_prompt
75
tokens_completion
792
native_tokens_prompt
73
native_tokens_completion
708
native_tokens_completion_images
(null)
native_tokens_reasoning
0
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
0.000182475
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788356195-e7LnhWUw1Z8luRfOg1gc"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788356195-oJcoG1M3U8zBPchy7kf7"
upstream_id
"2026090221363516037daeb88f4363"
provider_responses
0
endpoint_id
"8e9fe48b-2f91-41c3-a8a7-e4a93a8c4ff0"
id
"2026090221363516037daeb88f4363"
is_byok
false
latency
1847
model_permaslug
"z-ai/glm-5.3-flash-20260826"
provider_name
"Z.AI"
status
200
total_cost
0.000182475
cache_discount
(null)
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
35.00%
Matches word count
n/a
neededClean
false
words
518
0.00%
Dialogue to Total Word Ratio
Ratio: 74.52%, Deviation: 44.52%
neededClean
false
wordsTotal
522
wordsDialogue
389
17.5011%