NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-02-01M1GR6XP45530SK6SFMMB6F7B
Run Details
Model:
Z.AI GLM 5.3 (Reasoning, Max)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
4m 13s
Cost:
$0.067488
Provider:
Z.AI
Evaluation summary
Score
Evaluator
Details Link
99.98%
Matches word count
(details)
86.84%
Dialogue to Total Word Ratio
(details)
93.4135%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 70% 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 slapped a napkin down on the booth table and smoothed it flat with both palms, the way generals smooth maps before losing wars. "Recon complete. The Henderson party starts at two. Bouncy castle at two-fifteen. Magician at three. Presents at three-thirty, sharp." Bill leaned in and squinted. "This is a Denny's napkin." "It's a floor plan." "There's syrup where the bouncy castle should be." "That IS the bouncy castle. Keep up." She tapped a grease-stained corner. "We enter through the side gate during the cake song. Nobody moves during the cake song, Bill. Every adult in that yard becomes a hostage to propriety. It's our window." "And the dog?" Her shoulders dropped. "There's a dog." "There's always a dog." "Biscuit. Golden retriever. Sweet as pie until you get within six feet of that gift table. Then he becomes Cerberus with a wagging tail." "How do you know that?" "Additional recon." "You said recon was complete." "The recon was complete. The additional recon was additional." A waitress refilled their coffees, studied the napkin, and wisely chose not to ask. Bill rubbed his jaw. "And the clown?" Molly went pale. "Mr. Twisters. An unknown variable. He made a balloon octopus with moving arms, Bill. Moving arms. Clowns are unpredictable." "Nobody's ever outsmarted me because of a clown." "You've done two jobs." "Both clownless. That's a winning streak." Molly's expression suggested she was recalculating her cut. "One more concern. You go soft at children's parties." "I do not." "The Stringer job." "We agreed never to discuss the Stringer job." The Stringer job went un-discussed for a full three seconds. "You gave the presents back, Bill. You re-wrapped them. Better than the originals." "Their wrapping paper was a war zone, Molly. A child deserves clean lines on her birthday." "Getaway vehicle?" "The van's repainted. It says Balloon Delivery on the side." She frowned. "A clown will notice a balloon van." "That's hiding in plain sight, Molly." "That's walking into a rival's arms." He waved her off. "One more thing. The birthday boy would recognize me." "Why would an eight-year-old recognize you?" He examined his coffee very carefully. "I coach his soccer team." "That's not ideal." "And his T-ball team. His mother calls me Uncle Bill at the grocery store." Molly slowly set down her fork. "You're THE Uncle Bill? The Uncle Bill from the Hendersons' Christmas card?" "Which is why this job is already solved. I RSVP'd for both of us. You're my plus-one. I told Debra you do face painting." "I don't do face painting." "You'll learn by Saturday. It's a low-skill hustle. You have steady hands and no conscience. That's the whole résumé." She stared at the napkin, at the syrup castle, at the grease-stained gate. Somewhere out there, an eight-year-old was about to have the worst birthday in recorded history, and she was going to be there holding a paintbrush. "Fine," she said. "But no spiders. Spiders take too long." "Every great artist has limits," Bill said, and ordered pancakes to celebrate.
created_at
"2026-09-02T09:46:40.716Z"
9/2/2026, 9:46:40 AM
model
"z-ai/glm-5.3-20260816"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2869
moderation_latency
(null)
generation_time
253127
tokens_prompt
75
tokens_completion
12027
native_tokens_prompt
73
native_tokens_completion
15315
native_tokens_completion_images
(null)
native_tokens_reasoning
14594
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
0.0674882
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.19.0; linux; x64))"
http_referer
(null)
request_id
"req-1788342400-EZFw2ZqglW7sRmb1WzFP"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1788342400-rm25hE1B2A4Df1gvraew"
upstream_id
"20260902174640e7e4e562ee8d432d"
provider_responses
0
endpoint_id
"db975786-e7f6-4838-9e56-3e0b77f1bdc2"
id
"20260902174640e7e4e562ee8d432d"
is_byok
false
latency
2869
model_permaslug
"z-ai/glm-5.3-20260816"
provider_name
"Z.AI"
status
200
total_cost
0.0674882
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
99.98%
Matches word count
n/a
neededClean
false
words
502
86.84%
Dialogue to Total Word Ratio
Ratio: 71.68%, Deviation: 1.68%
neededClean
false
wordsTotal
512
wordsDialogue
367
93.4135%