NC Bench
Overview
About
Tests
Models
Model
Z.AI GLM 4.7
Test
Dialogue tags
Scenario
Write 500 words with 70% dialogue
Run ID
2026-02-08-01KGYXJHQ79ZPHXW6KNDKQSEXC
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 adjusted the wire in her ear, crouching low behind the particularly dense rhododendrons. "Target acquisition?" she whispered. "Visuals confirmed," Bill replied, checking his watch. "The piñata is swinging at three degrees east. The bounce castle is fully inflated. Air pressure looks critical." "And the package?" "The mountain of presents? It’s beautiful. Shiny, sparkly, untouched. I see at least three unboxed gaming consoles and what looks like a pony. An actual live pony." "We skip the pony. It’ll ruin our exit velocity." "Right. No livestock. You think the lock on the back gate is rigged?" "I saw a Dad eyeballing it earlier. Probably got a booby trap. A tripwire connected to a sprinkler system." "Wet bandits scenario? That’s amateur hour." "Don’t get cocky. This is suburbia, Bill. These people have HOAs, they have passive-aggressive notes, and they have zero tolerance for trespassers. What’s your exit strategy?" "The minivan is idling around the corner. I’ve got the booster seats ready for transport." "Booster seats?" "We want to blend in, don't we? If we get pulled over with a sack of toys, we look like thieves. If we look like we’re just driving our own kids home, nobody bats an eye." "That’s actually brilliant." "I’m not just a pretty face. So, I grab the loot, you stun the juggler?" "I don’t pay you to stun jugglers, I pay you to grab. We hit the pile when they light the candles. The flash from the match will blind the security cameras." "What about the birthday boy? He looks like a biter." "Bring pacifiers. Just in case." "Pacifiers. Check." "Remember, we want the high-value items. No clothes. No socks. If it’s itchy or educational, leave it behind." "Understood. STEM toys only." "And Bill?" "Yeah?" "If anyone sees you... cry." "Cry?" "Like a little girl. Tell them you lost your balloon. It’s the ultimate camouflage." "I hate this job sometimes." "You love the payday. Move out." "Moving."
created_at
"2026-02-08T15:25:14.745088+00:00"
Sunday, February 8th, 2026 at 3:25:14 PM GMT+00:00
model
"z-ai/glm-4.7-20251222"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
5734
moderation_latency
(null)
generation_time
138667
tokens_prompt
61
tokens_completion
2352
native_tokens_prompt
66
native_tokens_completion
2362
native_tokens_completion_images
(null)
native_tokens_reasoning
2054
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
0.00413942
router
(null)
provider_responses
0
id
"chatcmpl-R5HFAdwdpSAtLEtDYQbdgQZG"
status
200
is_byok
false
latency
66
endpoint_id
"ab79ffa8-8dcc-41c4-9a9f-6cc408135243"
provider_name
"DeepInfra"
model_permaslug
"z-ai/glm-4.7-20251222"
api_type
"completions"
id
"gen-1770564175-0ET7ScqFBgnTbsKQl5w8"
upstream_id
"chatcmpl-R5HFAdwdpSAtLEtDYQbdgQZG"
total_cost
0.00413942
cache_discount
0.00002048
upstream_inference_cost
0
provider_name
"DeepInfra"
Evaluation details
Result
Evaluator
Details
Meta Data
0.0000%
Matches word count
n/a
neededClean
false
words
324
0.0000%
Dialogue to Total Word Ratio
Ratio: 93.52%, Deviation: 23.52%
neededClean
false
wordsTotal
324
wordsDialogue
303
0.0000%