NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-24-01M39XH1Y1ATXTCJ69QQRBYFEN
Run Details
Model:
Aion 3.5 Mini (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
5m 24s
Cost:
$0.026584
Provider:
AionLabs
Evaluation summary
Score
Evaluator
Details Link
99.92%
Matches word count
(details)
1.66%
Dialogue to Total Word Ratio
(details)
50.7912%
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
The minivan sat across from the Hendersons' backyard, where a dragon-shaped bounce castle rippled and forty children shrieked like it was the apocalypse. Molly counted presents through binoculars. Bill ate chips. "Tell me what you see," she said. "A bouncy castle," Bill said. "Forty kids. One clown." "The clown's a problem." "It's a clown, Molly. I've fought clowns before." "You've never fought anything. You got winded beating a vending machine." "That machine owed me two dollars." Bill brushed crumbs off his shirt. "Focus. Presents are on the table by the garage. Dinosaur paper. I count thirty." "Thirty presents. For one kid?" "Rich people, Bill. They buy their children affection." "Can't buy their way out of getting robbed, though." Molly tapped the case file on her knee. It read: OPERATION PIÑATA. "We go in during the piñata. Eyes go up, we grab the table, out in ninety seconds." "What if the clown sees us?" "The clown is busy making balloon animals." "What if a kid sees us?" "Kids see us all the time. We're forgettable." "That's hurtful, but fine. What's my role?" "You're the lookout. You stand by the bouncy castle in the golden retriever costume." Bill stopped chewing. "I'm not being a dog at a child's birthday party." "It's the perfect cover. Dogs at parties are invisible. Kids will pet you. It's camouflage." "It's a fursuit, Molly." "It's a disguise." She dangled the head between the seats. Googly eyes. A tongue sewn to the zipper. "There's a tail. There's a big, floppy tail." "The tail sells it." Bill stared at the crumpled suit. "This is a new low. We've robbed museums, banks, a casino owned by a literal warlord." "And now a seven-year-old named Brayden." "Brayden?" "That's the birthday boy. He's getting a Nintendo Switch, a bike, and something heavy enough to rattle the table. Probably a pony." "We can't fence a pony." "We're not fencing anything. Presents are untraceable. No serial numbers. No guards." "Because they're guarded by parents." "Parents are distracted. It's a party. They're drinking mimosas and pretending to like each other." A clown emerged from behind the slide, twisting a balloon swan. Bill flinched. "What about the kid? He's going to be sad, Molly. A little boy, crying over stolen presents." "You robbed a nativity scene in Bethlehem, Pennsylvania." "That was different. That was Jesus. Jesus forgives." "Brayden will forgive too. Kids are resilient. Besides, half these gifts are regifts. I saw a candle box." "Stealing from children, Molly." "Stealing from a homeowner's insurance policy, technically. Parents will replace everything. Brayden ends up with double. We're basically a charity." "That's the worst math I've ever heard." "It's the best math. Tail on, bounce into position, and when the piñata comes out, signal me." "How?" "Bark twice." Across the lawn, a llama piñata was being hoisted into an oak. Bill sighed and pulled on the floppy golden head. Forty children shrieked with delight. "How do I look?" "Like a dog with a mortgage." He barked twice. Molly was already moving.
created_at
"2026-09-24T14:36:18.253Z"
9/24/2026, 2:36:18 PM
model
"aion-labs/aion-3.5-mini-20260923"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1382
moderation_latency
(null)
generation_time
324257
tokens_prompt
75
tokens_completion
15898
native_tokens_prompt
65
native_tokens_completion
18980
native_tokens_completion_images
(null)
native_tokens_reasoning
15094
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
0.0265837
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1790260578-qHqdl8TpEbrrO3P4fht9"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790260578-ivs8Z7HUOy2Dj8ny29q8"
upstream_id
"d055c9410be64c3180e179392c5604f4"
provider_responses
0
endpoint_id
"60ac3bdc-90ec-4769-9193-2c00b4d79562"
id
"d055c9410be64c3180e179392c5604f4"
is_byok
false
latency
1382
model_permaslug
"aion-labs/aion-3.5-mini-20260923"
provider_name
"AionLabs"
status
200
total_cost
0.0265837
cache_discount
0.0000338
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
99.92%
Matches word count
n/a
neededClean
false
words
503
1.66%
Dialogue to Total Word Ratio
Ratio: 79.05%, Deviation: 9.05%
neededClean
false
wordsTotal
506
wordsDialogue
400
50.7912%