NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-04-24-01KQ09XCBWD36RWP6PMSVPX822
Run Details
Model:
Z.AI GLM 5.1
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
6m 15s
Cost:
$0.086337
Provider:
Z.AI
Evaluation summary
Score
Evaluator
Details Link
100.00%
Matches word count
(details)
99.36%
Dialogue to Total Word Ratio
(details)
99.6821%
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 tapped the wooden pointer against the crude whiteboard diagram, her eyes narrowing at the brightly colored stick figures surrounding a poorly drawn bouncy castle. The tension in the dimly lit room was palpable. Bill sat on the sagging couch across from her, shoveling stale Cheerios into his mouth, looking entirely too relaxed for the sheer gravity of the upcoming operation. "This isn't a game. We are looking at Fort Knox in primary colors." Bill crunched loudly, his gaze drifting from the blueprint to the window as he calculated the odds. He swallowed hard, wiping a rogue drop of milk from his chin with a rugged sleeve. "It is a kid's party, Mol. How much security can a six-year-old have?" Molly threw her hands up in sheer exasperation, pacing frantically in front of the cluttered coffee table. She kicked a discarded soda can away. Her mind raced through the countless variables of their impending heist. "Have you ever tried taking a toy from a toddler? They have grip strength like a python and zero negotiation skills. There is a balloon animal guy. Those guys are always former military." Bill paused his chewing, finally contemplating the terrifying prospect of an ex-Navy SEAL twisting a squeaky poodle out of latex. The red circle around the presents table suddenly looked far less like a target and much more like a bullseye painted squarely on their backs. "We go in quiet. I distract the clown, you grab the goods." Molly shook her head violently, adjusting the tiny communicator nestled in her ear canal. The clown was an unpredictable variable that could completely derail the entire mission if left unchecked. She reached into her tactical duffel bag and triumphantly pulled out a massive, industrial-sized bag of gummy bears. "Is that the hyper-active blend? The ones that got banned in three states?" Molly grinned, a terrifying glint of mischief shining in her dark eyes. "Exactly. I dump this into the punch bowl. While the parents deal with sugar riots, we move on the loot." Bill rubbed his chin, visualizing the chaotic operation unfolding perfectly in his mind. Sneaking past the adults would be simple; they were usually far too busy drinking boxed wine in the kitchen to notice seasoned rogues slipping through the back door. "What about the piñata? If it breaks, they swarm." Molly pointed the wooden stick at the map once more, tapping a specific corner of the backyard with decisive, unyielding authority. "You are on piñata duty. Make sure the candy falls opposite the gifts. Draw fire." Bill stood up from the sofa, adjusting his collar with a renewed sense of profound patriotic duty. "I can do that. I have been practicing my swing." Molly winced visibly, remembering the catastrophic incident at the neighborhood Fourth of July barbecue. "Do not take out the patio furniture. We are thieves, not vandals." Bill nodded solemnly, offering a crisp, formal salute. They were true professionals, after all, or at least they were about to become legends.
created_at
"2026-04-24T17:51:58.857Z"
4/24/2026, 5:51:58 PM
model
"z-ai/glm-5.1-20260406"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2690
moderation_latency
(null)
generation_time
375237
tokens_prompt
77
tokens_completion
17514
native_tokens_prompt
66
native_tokens_completion
19601
native_tokens_completion_images
(null)
native_tokens_reasoning
18977
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
0.0863368
router
(null)
provider_responses
0
endpoint_id
"4ce9a513-17dc-4130-8960-2c4ceed3ea56"
id
"20260425015159fd62851c27b540d4"
is_byok
false
latency
2685
model_permaslug
"z-ai/glm-5.1-20260406"
provider_name
"Z.AI"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.1; linux; x64))"
http_referer
(null)
request_id
"req-1777053118-Wqt3Y1UUHdoHIRRCcEJ0"
session_id
(null)
api_type
"completions"
id
"gen-1777053118-YdRvV7jUCRm7zLipCuGC"
upstream_id
"20260425015159fd62851c27b540d4"
total_cost
0.0863368
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Z.AI"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
100.00%
Matches word count
n/a
neededClean
false
words
500
99.36%
Dialogue to Total Word Ratio
Ratio: 30.36%, Deviation: 0.36%
neededClean
false
wordsTotal
504
wordsDialogue
153
99.6821%