NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-02-01M1GWEPJ5N111FZVXN2XZCZMR
Run Details
Model:
Z.AI GLM 5.3 (Reasoning, Max)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
5m 45s
Cost:
$0.077522
Provider:
Z.AI
Evaluation summary
Score
Evaluator
Details Link
99.98%
Matches word count
(details)
92.05%
Dialogue to Total Word Ratio
(details)
96.0172%
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 planning session began, as all of Molly's planning sessions did, with a blueprint drawn on the back of a Chinese takeaway menu. Bill studied it from across the kitchen table and felt his soul leave his body slightly. It was a floor plan of number fourteen, Primrose Close, with the paddling pool labeled "MOTHER'S RUIN" for reasons he chose not to investigate. "So this is it," he said. "The big one. A child's birthday party. In Surrey." "Seven-year-old's birthday party," Molly corrected, tapping the conservatory with a biro. "Forty guests. Bouncy castle. Professional clown. Presents on the table by the French doors. That's our vault, Bill. Two grand in LEGO alone." Bill helped himself to a bourbon biscuit and chewed it with the slow deliberation of a man buying time. Molly took his silence as agreement and pressed on. She had reconnoitered the house for three days, which in Molly's language meant she had walked past it twice with a dog she did not own. She had identified the sight lines and the target's one strategic weakness. "The grandmother," she announced, "is the only credible threat. Eyes like a hawk. So we spike her sherry." "We are not drugging an old woman at a children's party, Molly." "Fine. We give her a second sherry. She'll do the rest herself." Bill spread his hands, searching for the version of this conversation where he won. There wasn't one. There never had been. According to the legend, he would enter through the side gate at two o'clock disguised as a caterer, in a polo shirt Molly had purchased, carrying a tray of sausage rolls she refused to explain. "Why forty sausage rolls?" he asked. "Because nobody questions a man with a tray." He could feel the logic of it settling over him like a warm, terrible blanket. It was, infuriatingly, a decent plan. The parents would be in the garden arguing about charcoal. The clown would have the children in a conga line. The dog was elderly, deaf, and, according to Molly's notes, "morally flexible." "And the exit?" he said. "Same as the entrance, only with a cool box full of other people's presents and a spring in your step. Rich by teatime." Bill closed his eyes. He thought about the Calais job. He thought about prison. He thought, briefly and longingly, about an honest life. Then he thought about the ride-on jeep, and how much his nephew would love it, and how nobody would miss one jeep among forty. "There's a bouncy castle," Molly added, watching him the way a hawk watches a sherry glass. "It would be a shame not to use it." "That's not a tactical opinion," said Bill. "It could be a tactical distraction," said Molly, far too innocently. "It's your birthday present to yourself, and we both know it." Molly said nothing, which in Molly's language was a confession. Bill sighed the sigh of the professionally doomed, picked up a practice sausage roll, and began work on the timing.
created_at
"2026-09-02T11:00:49.872Z"
9/2/2026, 11:00:49 AM
model
"z-ai/glm-5.3-20260816"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1019
moderation_latency
(null)
generation_time
344929
tokens_prompt
75
tokens_completion
16124
native_tokens_prompt
73
native_tokens_completion
17612
native_tokens_completion_images
(null)
native_tokens_reasoning
16957
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
0.07752204
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788346849-tmfNmcDNm4Si4Pu4LymW"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788346849-0tdHNUstPT6ym3uYT6hP"
upstream_id
"2026090219004955bce8b7b60e4a5c"
provider_responses
0
endpoint_id
"db975786-e7f6-4838-9e56-3e0b77f1bdc2"
id
"2026090219004955bce8b7b60e4a5c"
is_byok
false
latency
1019
model_permaslug
"z-ai/glm-5.3-20260816"
provider_name
"Z.AI"
status
200
total_cost
0.07752204
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
99.98%
Matches word count
n/a
neededClean
false
words
502
92.05%
Dialogue to Total Word Ratio
Ratio: 31.29%, Deviation: 1.29%
neededClean
false
wordsTotal
505
wordsDialogue
158
96.0172%