NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-24-01M39VEQXMG00P5SCYDVV46TQT
Run Details
Model:
Aion 3.5 Mini (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
32.1s
Cost:
$0.001460
Provider:
AionLabs
Evaluation summary
Score
Evaluator
Details Link
6.09%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
3.0453%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 50% 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 pressed binoculars to her face and squinted through the picket fence at the backyard chaos beyond. "There it is, Bill. The score of a lifetime. Forty presents, minimum, wrapped in that shiny paper with the dinosaurs on it." Bill adjusted his ski mask, which had a cartoon cat printed on it because it was the only one the party store had left. "I still don't understand why we're doing this. These are children's presents." "Exactly. Nobody guards children's presents. No lasers, no vaults, no attack dogs. Just a folding table and a banner that says Happy Birthday." "That banner is load-bearing evidence, Molly." "It's a banner, Bill." Bill chewed his lip. "What about the clown?" Molly lowered the binoculars slowly. "The clown?" "Yeah. Bubbles. He's been making balloon animals for two hours straight. That man hasn't blinked. I think he's ex-military." "He is not ex-military. He's a guy in suspenders named Bubbles." "Nobody's naturally named Bubbles, Molly. That's a codename. He's clearly running some kind of operation." "Fine. We neutralize the clown. You distract him, I grab the loot." "How do I distract a clown?" "Ask him how he got into clowning. Clowns love that." Bill nodded, writing it on his palm. "Okay. And the piñata?" "What about it?" "It's hanging from the oak tree like it's taunting us. That's where they keep the good candy, right? The king-size bars?" "That's the reserve stash, yes." "Then the piñata is the safe. You don't crack a safe at the beginning of a heist, Molly. You crack it at the end, after the dramatic twist." Molly stared at him. "Bill, it's a papier-mâché donkey." "It's a papier-mâché donkey with a secret, and we both know it." She sighed and checked her watch. "Party ends at four. Parents start collecting kids at three-thirty. We go in at three when everyone's hopped up on juice and unable to focus on anything but cake." "And the pony?" "The what?" "The pony. There's a pony giving rides in the side yard. That's a patrol animal, Molly. That's a mounted unit." "It's a birthday pony, Bill. It's named Cupcake. It weighs less than the piñata." "That's exactly what a trained infiltration pony would want us to think." Molly rubbed her temples. "New plan. You stand at the fence and worry about the pony. I take the presents." "That's not a team heist. That's a solo heist with emotional support." "You're the emotional support, Bill. You're support and you're emotional. Now, hand signals. Fist means stop. Open palm means go. Two fingers means clown incoming." "What does a thumbs up mean?" "Nothing. Never give me a thumbs up. You'll blow the whole thing." Bill gave her a thumbs up. "That's it. We're going in before you learn any more gestures." They vaulted the fence. Immediately, Bill stepped on a squeaky toy. Then another. Then twelve more, in a chain reaction that sounded like a flock of dying rubber ducks. "Bubbles heard that," Bill whispered. "Bubbles is making a balloon giraffe. Walk faster." A shadow fell over them. It was the pony. "Bill, don't panic." "I'm not panicking, Molly. I'm negotiating."
created_at
"2026-09-24T14:00:05.525Z"
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
4653
moderation_latency
(null)
generation_time
31810
tokens_prompt
75
tokens_completion
1087
native_tokens_prompt
65
native_tokens_completion
1010
native_tokens_completion_images
(null)
native_tokens_reasoning
287
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
0.0014595
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.21.0; linux; x64))"
http_referer
(null)
request_id
"req-1790258405-s2m0Jb6xF9Mywj7Qlpc5"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790258405-sE1z57lCwAhB3imO8hnh"
upstream_id
"b23b2107845a4b6193ab4a431e917136"
provider_responses
0
endpoint_id
"60ac3bdc-90ec-4769-9193-2c00b4d79562"
id
"b23b2107845a4b6193ab4a431e917136"
is_byok
false
latency
4622
model_permaslug
"aion-labs/aion-3.5-mini-20260923"
provider_name
"AionLabs"
status
200
total_cost
0.0014595
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
6.09%
Matches word count
n/a
neededClean
false
words
523
0.00%
Dialogue to Total Word Ratio
Ratio: 77.84%, Deviation: 27.84%
neededClean
false
wordsTotal
528
wordsDialogue
411
3.0453%