NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-23-01M37P9X523P4VWQ5077GQR4JG
Run Details
Model:
Aion 3.5 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
23.9s
Cost:
$0.005809
Provider:
AionLabs
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
0.0000%
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 crouched behind the hedge of number fourteen Primrose Lane, binoculars pressed to her face, and exhaled slowly. "There," she whispered. "Bouncy castle at two o'clock. Clown by the gazebo. And a table of presents that would make a museum curator weep." Bill wriggled up beside her, his camouflage trousers already stained with mud from the park. He'd been in the game twenty years—banks, casinos, one memorable incident involving a crown jewels replica—and yet here he was, squinting at a garden full of six-year-olds like it was Fort Knox. "I don't like it," he said. "There's a petting zoo. That's a variable we cannot control." "It's a pony, Bill. One pony. And a goat, apparently." "Goats are chaos incarnate." Molly lowered the binoculars and fixed him with the stare that had once made a Monaco security chief hand over his access card. "The Henderson party has forty-three guests. Forty-three presents. I counted the gift table at drop-off this morning when I posed as the balloon delivery woman. There's a life-sized robot dinosaur in there, Bill. Retail value, three hundred quid. Plus a telescope. Plus something in wrapping paper with actual gold leaf on it." Bill rubbed his jaw. In his experience, the most dangerous heists weren't the ones with lasers and guards—they were the ones with emotional variables. And nothing on earth carried more emotional variables than a child's birthday party. "The parents," he said. "What's the parent situation?" "Mother's inside doing food prep. Father's in the gazebo with a lager and a clipboard. He's wearing a novelty tie, which means his situational awareness is compromised." "Novelty tie is a good sign," Bill admitted. "A man in a novelty tie has surrendered to the day." "Exactly. Now here's the plan." Molly pulled a napkin from her pocket, upon which she had drawn a diagram that looked, to the untrained eye, like a spider had died in ketchup. "We go in as the entertainers. The magician cancelled this morning—food poisoning, I may have had something to do with that—and they're desperate. I've already volunteered us. You're doing balloon animals. I'm doing face painting." Bill stared at her. "I don't know how to make balloon animals." "You're doing balloon animals badly, Bill. It's a children's party. They have no standards. Twist it into something vaguely dog-shaped and they'll applaud until their arms fall off." "And the presents?" "Phase two. During the cake ceremony—every eye in that garden will be on the cake, trust me, it's a three-tier monstrosity with a fondant dragon— I slip the presents into the wheelie bin we've stashed by the side gate. You cover me by 'accidentally' releasing the balloons." "All of them?" "All of them. Total sensory chaos. Parents will be weeping. Children will be screaming. It'll be beautiful." Bill was quiet for a moment, watching a small child in a dinosaur costume tackle another small child in a dinosaur costume near the trampoline. "There's something you should know," he said. "About the goat." Molly narrowed her eyes. "What about the goat?" "The petting zoo man. I spoke to him this morning. Chatty fellow. Turns out the goat's name is Bandit." He paused for dramatic effect, a skill honed over decades. "And Bandit has a history of eating anything that isn't nailed down. Including, on one occasion, a magician's entire bag of tricks." Molly's face went pale. "You're saying we have a rival operative in the field." "I'm saying, Molly, that we might be out-thieved by a goat." She looked back through the binoculars at the gift table gleaming in the afternoon sun, then at Bandit, who was already eyeing the fence with what could only be described as criminal intent. "New plan," she said grimly. "We move up the timeline. We go in during pass-the-parcel." Bill swallowed. "Pass-the-parcel is sacred, Molly." "So is a life-sized robot dinosaur. Saddle up. And Bill—grab the good wrapping paper first. The gold leaf one. That one's mine."
created_at
"2026-09-23T17:51:35.091Z"
9/23/2026, 5:51:35 PM
model
"aion-labs/aion-3.5-20260923"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
913
moderation_latency
(null)
generation_time
23808
tokens_prompt
75
tokens_completion
1066
native_tokens_prompt
65
native_tokens_completion
960
native_tokens_completion_images
(null)
native_tokens_reasoning
50
native_tokens_cached
65
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
0.00580875
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.20.0; linux; x64))"
http_referer
(null)
request_id
"req-1790185895-Gp2gY9DsWy1mvZetEyyj"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790185895-ZC5GAFVlf5sdXrT5oUMl"
upstream_id
"012ebc7780974e2489a025a517fb0a7f"
provider_responses
0
endpoint_id
"09b1fc56-3888-42bd-b7a0-926851c7f9d6"
id
"012ebc7780974e2489a025a517fb0a7f"
is_byok
false
latency
913
model_permaslug
"aion-labs/aion-3.5-20260923"
provider_name
"AionLabs"
status
200
total_cost
0.00580875
cache_discount
0.00014625
upstream_inference_cost
0
provider_name
"AionLabs"
response_cache_source_id
(null)
data_region
"global"
workspace_id
"97e315e5-d303-487d-83c1-83180e8a13d4"
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
664
0.00%
Dialogue to Total Word Ratio
Ratio: 62.78%, Deviation: 32.78%
neededClean
false
wordsTotal
677
wordsDialogue
425
0.0000%