NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-23-01M37Q9CEZBNXGJXT2TVZ3C6HP
Run Details
Model:
Aion 3.5 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
35.8s
Cost:
$0.007465
Provider:
AionLabs
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.03%
Dialogue to Total Word Ratio
(details)
0.0149%
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 Hendersons' hedge, binoculars pressed to her face, while Bill unrolled a diagram on the lawn with the gravity of a general reviewing battle plans. The diagram was drawn in crayon. There was a smiley sun in the corner. "Okay," Bill whispered, tapping the paper with a pencil. "The perimeter. What are we looking at?" "A fence," said Molly. "It's three feet tall. It's decorative. A determined toddler could clear it." "But could a determined adult clear it?" "Bill, I could step over it. You might have to do the thing where you lean on it and swing your leg." Bill nodded gravely and made a note. Behind the fence, the party roared on — twelve six-year-olds, a bouncy castle pulsating like a diseased lung, and a table of presents stacked in a glittering tower that made Molly's professional heart go pitter-pat. Somewhere, a child was screaming with what sounded like joy but could easily have been something worse. "The presents are the easy part," Bill said. "It's the security we need to worry about." He pointed his pencil at a figure near the grill. "That's the father. He's armed." "He's holding a spatula." "He's holding a spatula *menacingly*." Molly lowered her binoculars and looked at him. Bill had been in the game for twenty years. He'd tunneled into vaults, rappelled down elevator shafts, once spent three days inside a decorative horse. And yet every job, without fail, he found the one thing to be terrified of, and it was never the thing anyone else would pick. "There's a clown," Bill continued, voice dropping. "I saw him earlier. He makes animals out of balloons. Do you know what that means?" "Whimsy?" "Distraction techniques. Misdirection. That's not a clown, Molly. That's a fellow professional. He's running his own operation right under our noses." Molly rubbed her eyes. Across the yard, the clown was indeed making a balloon poodle, which in Molly's professional assessment meant he was a clown making a balloon poodle. She'd been in the game eleven years and had learned that the difference between a hardened criminal and an idiot was mostly a matter of paperwork. "The bouncy castle," Bill said, jabbing the crayon diagram. "I've marked it as the laser grid. We go over it, not through it. If you touch the sides—" "You get yelled at by a man in a lawn chair." "You get *ejected* by a man in a lawn chair." Molly had to admit he wasn't entirely wrong about that part. She'd seen the man in the lawn chair. He had the dead-eyed vigilance of someone paid in pizza slices and gratitude, which was to say, someone with nothing left to lose. "Fine. We go over the castle. Then what?" "Then," Bill said, straightening to his full crouching height, "we hit the table. I take the big box with the rocket ship on it. You take everything else. We regroup at the piñata, which I've designated as the safe house." "You've designated a donkey full of candy as the safe house." "It's structurally sound and emotionally neutral." Molly looked at the presents, then at the fence, then at Bill, who was now applying face paint in the reflection of his own binoculars. It was, she reflected, the worst plan she'd ever been part of, which meant statistically it had an excellent chance of working. "On three," said Bill. "One. Two." "Wait — why is there a child in the hedge with us?" There was, in fact, a child in the hedge with them. She was holding a juice box and looking at Bill's face paint with an expression of profound pity. "Cool camouflage," she said. "My dad's a clown."
created_at
"2026-09-23T18:08:46.578Z"
9/23/2026, 6:08:46 PM
model
"aion-labs/aion-3.5-20260923"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
12838
moderation_latency
(null)
generation_time
35728
tokens_prompt
75
tokens_completion
1361
native_tokens_prompt
65
native_tokens_completion
1236
native_tokens_completion_images
(null)
native_tokens_reasoning
457
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
0.00746475
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.20.0; linux; x64))"
http_referer
(null)
request_id
"req-1790186926-Ppl9Q1PluvZCDNwFKgyB"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790186926-EhcZNxMzlOuxqEu3Zmba"
upstream_id
"84f7e77325c14cc28b462292b837dad1"
provider_responses
0
endpoint_id
"09b1fc56-3888-42bd-b7a0-926851c7f9d6"
id
"84f7e77325c14cc28b462292b837dad1"
is_byok
false
latency
12838
model_permaslug
"aion-labs/aion-3.5-20260923"
provider_name
"AionLabs"
status
200
total_cost
0.00746475
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
616
0.03%
Dialogue to Total Word Ratio
Ratio: 42.74%, Deviation: 12.74%
neededClean
false
wordsTotal
620
wordsDialogue
265
0.0149%