NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-02-01M1GTV3PAGFFXK7VT017ABDEF
Run Details
Model:
Z.AI GLM 5.3 Flash (Reasoning, Max)
Test:
Dialogue tags
Scenario:
Write 200 words with 50% dialogue
Duration:
5m 35s
Cost:
$0.004296
Provider:
Z.AI
Evaluation summary
Score
Evaluator
Details Link
97.63%
Matches word count
(details)
99.79%
Dialogue to Total Word Ratio
(details)
98.7065%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 200 word scene that has 50% 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 unrolled the blueprint across the hood of the van, pinning the corners down with four garden gnomes she'd stolen on principle. Down the block, a mechanical pony whinnied on a loop, taunting them. "Target's the Wilson kid's birthday," she said. "Saturday. Backyard of 42 Maple." Bill squinted at the diagram. "Security?" "Two parents, one bouncy castle, and a clown." He went pale, reaching for his stress balloon. "Nobody said clown. I have a condition." "It's a condition of the job." "A phobia. Clown-related. Formative trauma." "You're fifty-three." He traced the fence line with a trembling finger. "What's the haul?" "Fifteen presents, minimum. Kid's popular. Plus a piƱata stuffed with twenties." "What is this family, cartel?" Molly nodded like this explained everything. "Pediatric dentists. Here's the play. You're the distraction." "I do safecracking. I do roofs. Not birthday clowns." "You juggle, the clown broods, I vacuum the gift table." "I've never juggled in my life." She patted his cheek like a boxing coach, then pressed three rubber duckies into his hands. "It's a children's party. Nobody expects technique." He weighed a duck. It squeaked. "Fine. But if he honks at me, I'm gone." Somewhere beyond the hedge, the bouncy castle inflated slowly, like a fat orange promise.
created_at
"2026-09-02T10:32:39.378Z"
9/2/2026, 10:32:39 AM
model
"z-ai/glm-5.3-flash-20260826"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
903
moderation_latency
(null)
generation_time
334733
tokens_prompt
75
tokens_completion
14503
native_tokens_prompt
73
native_tokens_completion
17176
native_tokens_completion_images
(null)
native_tokens_reasoning
16854
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
0.004295635
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788345159-b1N0srcZ1rIjdCsC8UDT"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788345159-LisnkNsUllLBphGneFiD"
upstream_id
"202609021832390e56d0c7cad24450"
provider_responses
0
endpoint_id
"8e9fe48b-2f91-41c3-a8a7-e4a93a8c4ff0"
id
"202609021832390e56d0c7cad24450"
is_byok
false
latency
903
model_permaslug
"z-ai/glm-5.3-flash-20260826"
provider_name
"Z.AI"
status
200
total_cost
0.004295635
cache_discount
0.00000768
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
97.63%
Matches word count
n/a
neededClean
false
words
207
99.79%
Dialogue to Total Word Ratio
Ratio: 52.15%, Deviation: 2.15%
neededClean
false
wordsTotal
209
wordsDialogue
109
98.7065%