NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-24-01M39VEQXE27KC7JWQHJKD2RDM
Run Details
Model:
Aion 3.5 Mini (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
4m 14s
Cost:
$0.014314
Provider:
AionLabs
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
96.80%
Dialogue to Total Word Ratio
(details)
48.4002%
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 van smelled like stale cheese puffs and regret. Rain hammered the roof, blurring the world outside into a watercolor smear of gray. Across the street, the Sunnyvale Community Center glowed with aggressive yellow light. A cardboard banner taped to the double doors read: HAPPY 7TH BIRTHDAY, BRADLEY! in letters that were aggressively cheerful. Molly adjusted her binoculars, watching a stream of minivans unload a cargo of small, screaming humans. The plan was simple, elegant, and exactly the kind of thing that got people arrested in Newark. She had drawn it on the back of a Denny's placemat, complete with arrows and a stick figure representing Bill looking 'non-threatening.' Beside her, Bill wrestled with a juice box he'd found under the seat. He stabbed the straw in anyway, hoping for a miracle. He looked at the diagram, then at the community center, then back at the diagram. "I still say we're out of our depth. These are children. What if they scream? What if they have super soakers?" Molly lowered the binoculars. She explained that the children would be hopped up on cake and fruit punch, possessing the attention span of a goldfish. They just needed nine minutes of goldfish focus. The real issue, she noted, was the logistics. Thirty pounds of shrink-wrapped plastic wasn't going to walk itself to the getaway car. "Ghosts with shopping carts? Because that's the only way we're hauling thirty pounds of plastic." Molly pointed to a red Radio Flyer wagon tucked behind a spare tire. It was perfect. It blended in. It said 'I belong here. I am a fun uncle.' Bill thought it said 'I am a child abductor,' but he kept that to himself. "What about the magician? What if the rabbit bites me?" Bill nodded toward a sign featuring a man with a terrifying, unblinking smile. The Amazing Marco. Molly explained that Marco was a hack who did card tricks. The real security was the bouncy castle, situated right next to the present pile like an inflatable fortress. "Marco is a hack. He does card tricks. The real obstacle is the bouncy castle." They would have to wait for the piñata. That was when the chaos peaked. Everyone would be blindfolded, swinging a bat at empty air. That was their window. "So we wait for the piñata. Chaos peaks. We grab and go." Bill stared out the window at a group of kids chasing a dog. He felt a familiar twinge of guilt, the same one he'd felt when they'd stolen the collection plate from St. Mary's, or the donation jar from the animal shelter. "I just feel like we're hitting a new low. Stealing from a seven-year-old." Molly reminded him that they were liberating excess consumer goods from a household that clearly spoiled their child. The kid had a trust fund. "We're liberating consumer goods. He won't miss a Nerf blaster." "It's an Xbox, Molly." Molly checked her binoculars again, zooming in on the gift table. A large box in the center of the pile did seem to be shifting. A small, horned head poked through the bow. "Look at the ribbon on that last box. It's moving." Bill leaned forward, squinting through the rain-streaked glass. "That's not a puppy, Molly. That's a goat." Molly checked her notes. Petting zoo. Okay, new plan. "Petting zoo. Okay, new plan. Grab the electronics, skip the livestock, and we are out before the cake is cut." "And if the clown sees us? What if he juggles us? I don't want to be juggled, Molly." Molly went pale. She turned back to the window, her jaw tight. "We do not speak of the clown. The clown is a wildcard."
created_at
"2026-09-24T14:00:05.541Z"
9/24/2026, 2:00:05 PM
model
"aion-labs/aion-3.5-mini-20260923"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
4564
moderation_latency
(null)
generation_time
253814
tokens_prompt
75
tokens_completion
10160
native_tokens_prompt
65
native_tokens_completion
10192
native_tokens_completion_images
(null)
native_tokens_reasoning
9265
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
0.0143143
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1790258405-OnYCIo9AsEN9e4PURbMR"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790258405-quuNQa67TBPX7D66oopo"
upstream_id
"26b3ab7bd409403e83589b5d8a824c7b"
provider_responses
0
endpoint_id
"60ac3bdc-90ec-4769-9193-2c00b4d79562"
id
"26b3ab7bd409403e83589b5d8a824c7b"
is_byok
false
latency
4564
model_permaslug
"aion-labs/aion-3.5-mini-20260923"
provider_name
"AionLabs"
status
200
total_cost
0.0143143
cache_discount
(null)
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
615
96.80%
Dialogue to Total Word Ratio
Ratio: 30.81%, Deviation: 0.81%
neededClean
false
wordsTotal
620
wordsDialogue
191
48.4002%