NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-24-01M39WRC4N1WD68GB3CHRT0QMJ
Run Details
Model:
Aion 3.5 Mini (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
1m 5s
Cost:
$0.004314
Provider:
AionLabs
Evaluation summary
Score
Evaluator
Details Link
9.61%
Matches word count
(details)
0.84%
Dialogue to Total Word Ratio
(details)
5.2266%
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 van idled across the street from the community center, where a banner reading HAPPY 7TH BIRTHDAY TOMMY flapped cheerfully in the breeze. Molly lowered her binoculars and turned to Bill, who was eating a sandwich. "Okay. Rolling pinata at three o'clock, clown on stilts, and I count nine adults near the present table." "Nine?" Bill said. "That's a lot of chaperones." "That's a lot of witnesses." "Relax. We go in as balloon sellers." "There's already a balloon seller. There's a guy dressed as a hot dog. There's a woman making balloon animals inside of a balloon animal. This party has security I've never seen." "Then we go as clowns." "We are not going as clowns. I have a reputation." "A reputation for what? Last month we stole a shipment of novelty mugs." "Those mugs were rare. Limited edition. They squeak when you tip them." Bill set down his sandwich and pulled out a napkin covered in diagrams. "New plan. The presents get moved inside at two for the unwrapping. That's our window. You distract the parents, I grab the loot." "Distract them how?" "You're charming." "I'm a professional." "Professional charm. Tell them you're from a magazine. Everybody panics when a photographer shows up unannounced. They'll be scrubbing baseboards for an hour." Molly nodded slowly. "The Birthday Gazette. I like it. And you?" "I'm the cake guy." "There is a cake guy. He's wearing a chef's hat and crying because the fondant melted." "Then I'm the replacement cake guy. Nobody questions a second cake. Kids love cake. Two cakes is a jackpot." "Two cakes is suspicious, Bill. Two cakes says somebody dropped the first cake." "Fine, then I'm Tommy's uncle." "You don't know Tommy." "Seven-year-olds don't know their uncles either. That's the beauty of uncles. Any adult who shows up and says 'wow, you got big' gets believed." Molly stared at him for a long moment. "That's... actually not terrible." "I have nieces. I've been to parties. I know the terrain." "Fine. You're Uncle Bill, I'm the press. But what about the actual uncle? Brown sweater, by the gift table, holding a juice box he didn't pour." Bill raised the binoculars. "That's the dad." "That's the dad?" "He looks exhausted." "He's gripping that juice box like it owes him money. Okay, new problem. The dad knows the real uncles." "So we wait for him to fall asleep. Look at him. Forty minutes of quiet and he's out on that folding chair." "You want to wait out a dad?" "Dads at parties are furniture, Molly. He'll go down." Molly checked her watch, then checked it again, as if time itself might be planning against them. "Fine. At nap time, you're Uncle Bill, I'm the press, and we're out the fire exit with forty presents in under six minutes." "And the pinata?" "We leave the pinata." "But the candy inside—" "We are thieves, Bill, not monsters. Besides, that candy's a trap. It's how they get you to come back next year." Bill nodded solemnly and finished his sandwich. Outside, the clown on stilts wobbled past the window, and neither of them said a word.
created_at
"2026-09-24T14:22:49.505Z"
9/24/2026, 2:22:49 PM
model
"aion-labs/aion-3.5-mini-20260923"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
1182
moderation_latency
(null)
generation_time
65009
tokens_prompt
75
tokens_completion
3249
native_tokens_prompt
65
native_tokens_completion
3073
native_tokens_completion_images
(null)
native_tokens_reasoning
2463
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
0.0043139
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1790259769-HY38EN0NQHCVikBlFboK"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790259769-zEr4LkYgZIeHaTnKRUKV"
upstream_id
"f847fb61b58b4a4da8dc7bade9862726"
provider_responses
0
endpoint_id
"60ac3bdc-90ec-4769-9193-2c00b4d79562"
id
"f847fb61b58b4a4da8dc7bade9862726"
is_byok
false
latency
1182
model_permaslug
"aion-labs/aion-3.5-mini-20260923"
provider_name
"AionLabs"
status
200
total_cost
0.0043139
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
9.61%
Matches word count
n/a
neededClean
false
words
522
0.84%
Dialogue to Total Word Ratio
Ratio: 79.77%, Deviation: 9.77%
neededClean
false
wordsTotal
524
wordsDialogue
418
5.2266%