NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-03-15-01KKS2Q71QA9P8MHWPSK0ARVQ7
Run Details
Model:
Z.AI GLM 5 Turbo
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
1m 26s
Cost:
$0.049566
Provider:
Z.AI
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
100.00%
Dialogue to Total Word Ratio
(details)
100.0000%
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
The blueprints were spread across the stolen folding table, weighted down by a warm half empty juice box and a dry stale crustless sandwich. Bill adjusted his black ski mask, pushing the mouth hole up so he could breathe without fogging his glasses. It was ninety degrees outside, but a true professional never compromises on aesthetics. Molly sat across from him, intensely sharpening a plastic cake knife with a sharp metal nail file. She wore a heavy tactical vest filled with loud brightly colored party hats instead of standard issue grenades. "Target is a bouncy castle in the backyard. We go in through the ball pit, grab the loot from the picnic table, and escape before the clown notices." Bill tapped a chubby finger on the wrinkled laminated diagram. The inflatable structure looked like a garish pink dragon, breathing harmless foam rather than actual fire. Getting inside the castle would require immense agility, perfect precision, and a total disregard for personal dignity. The floor was notoriously unstable. "Guard rotation is tight. One clown on the patio. A magician doing card tricks by the punch bowl. Bunch of highly unpredictable toddlers on the swing set. Those little monsters are covered in sticky jam." Molly stopped filing the dull plastic and pulled a cheap pair of toy binoculars from her pocket. She had spent three grueling hours surveilling the premises from behind a prickly hedge, narrowly avoiding being hit by a stray water balloon on four separate occasions. The enemy was relentless and completely immune to reason. "Toddlers are the worst. They have no situational awareness, sticky fingers, and a complete lack of respect for personal boundaries. You trip over one, and suddenly you are crying." Bill grimaced, vividly remembering his last encounter with an angry five year old. The tiny criminal had somehow stolen his wallet while loudly demanding a grape juice pouch. It was a deeply humiliating defeat that still haunted him during long, sleepless nights. He quickly checked his pockets defensively. "I will create a diversion. I grab the gifts. You keep the magician busy." Molly nodded slowly, purposefully pulling a shiny red kazoo from her dark vest pocket. It was a tragic artifact from a previous job gone horribly wrong, but it still held terrifying acoustic potential. If she played it badly enough, the birthday clown would be forced to intervene. "Do not play the kazoo. Last time you used that wretched instrument, we ended up fleeing a retirement home in separate ambulances. This is a clean extraction. No casualties, no tears." Bill sighed heavily, slipping the modified plastic cake knife into his very heavy leather boot. It was significantly duller than a wooden spoon, but it was the principle of the entire thing. They carefully packed up the stolen blueprints, chugged the terrible stale coffee, checked their silly disguises, and prepared to face the incredibly dangerous suburban criminal syndicate waiting in the sunny backyard. "Just steal the loot and run. Please do not talk to any children."
created_at
"2026-03-15T15:44:30.271Z"
3/15/2026, 3:44:30 PM
model
"z-ai/glm-5-turbo-20260315"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2666
moderation_latency
(null)
generation_time
85869
tokens_prompt
77
tokens_completion
13519
native_tokens_prompt
66
native_tokens_completion
15485
native_tokens_completion_images
(null)
native_tokens_reasoning
14882
native_tokens_cached
64
num_media_prompt
(null)
num_input_audio_prompt
(null)
num_media_completion
0
num_search_results
(null)
origin
"https://ncbench.com/"
is_byok
false
finish_reason
"stop"
native_finish_reason
"stop"
usage
0.049566208
router
(null)
provider_responses
0
endpoint_id
"9ce3b132-3a17-4fea-8092-0e8e519ed32b"
id
"20260315234432314132e673bf41ca"
is_byok
false
latency
2666
model_permaslug
"z-ai/glm-5-turbo-20260315"
provider_name
"Z.AI"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.0; linux; x64))"
http_referer
(null)
api_type
"completions"
id
"gen-1773589470-YxVqXRNwKyLhbaXapwBd"
upstream_id
"20260315234432314132e673bf41ca"
total_cost
0.049566208
cache_discount
0.00006144
upstream_inference_cost
0
provider_name
"Z.AI"
Evaluation details
Result
Evaluator
Details
Meta Data
100.00%
Matches word count
n/a
neededClean
false
words
500
100.00%
Dialogue to Total Word Ratio
Ratio: 30.00%, Deviation: 0.00%
neededClean
false
wordsTotal
500
wordsDialogue
150
100.0000%