NC Bench
Overview
About
Tests
Models
Vendors
Compare
Run: 2026-02-24-01KJ8VKYN86GF01MP5EQQ9QB49
Run Details
Model:
Aion 2.0
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
32.6s
Cost:
$0.002178
Provider:
AionLabs
Evaluation summary
Score
Evaluator
Details Link
0.03%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
0.0152%
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
The van smelled of stale coffee and regret. Molly smoothed the blueprint across a cooler marked 'FISHING TACKLE'. Bill peered at it, his face illuminated by the ghostly glow of a battery-powered lantern. “This is the target?” Bill asked, his voice flat. “The Anderson residence,” Molly said, tapping the paper. “Saturday, two PM. The package will be in the main living area, concentrated around the base of a decorated fir tree.” “The ‘package’ being…” “Presents, Bill. Wrapped, untagged, and completely unsecured. Intel suggests a high concentration of plastic, some electronic devices, and a shocking amount of gift cards.” Bill rubbed his temples. “We’re planning a felony-grade extraction on a six-year-old’s birthday party.” “It’s a *dual* birthday. Twins. Double the haul. Look,” she said, pointing to a sketched rectangle. “Primary obstacle: the adults. They’ll be distracted by conversation, potato salad, and attempting to moderate sugar intake. Secondary obstacle: the marks themselves. Unpredictable, loud, but easily neutralized with a well-placed piñata.” “Neutralized?” “Distracted. Semantics. Now, entrance.” Molly produced a laminated photo of a colorful, inflatable castle. “We go in disguised as the bounce house delivery team. It gets us onto the property, no questions asked. We establish the bounce house *here*,” she pointed to the backyard, “creating a natural diversion. All kids will migrate to it. The living room will be clear.” Bill stared at the photo. “You want us to actually set up a bounce house?” “It’s our cover. I’ve practiced. It’s mostly just plugging in a loud fan.” “And the exit strategy? We can’t exactly sprint for the fence with a sack of toys over our shoulder.” “We don’t sprint. We *repurpose*. See this alley?” Her finger traced a line behind the property. “The bounce house company van—our van—parks here. We load the presents into empty bounce house bags. It’s just removing equipment. Clean, professional.” Bill leaned back, the van’s metal wall cold against his spine. “What’s the contingency for the dog?” Molly blinked. “Dog?” “There’s always a dog. A yappy little thing that goes for the ankles.” “The file didn’t mention a dog.” “The file,” Bill said, “is about a toddler’s party. It’s not exactly a CIA dossier. What about the cake? Sugar crashes lead to meltdowns. Meltdowns lead to parents gathering inside. That’s a hard lock on the package.” Molly considered this. “We time the extraction for the cake presentation. Maximum crowd focus is on candles, not us. As for the dog…” She reached into a duffel bag and pulled out a bulky, fur-covered object. “We bring our own ‘dog’. A convincing enough decoy to lure any real one away. I got it half-price after Halloween.” Bill took the limp, plush husky. One glass eye was missing. “It’s horrifying.” “It’s tactical,” Molly corrected. “So? Are you in? This could be the biggest score of the year. Low security, high reward.” Bill looked from the sad plush dog to the eager glint in Molly’s eyes, then to the blueprint of a suburban living room soon to be filled with brightly colored boxes. He sighed, a long, weary sound that spoke of fallen dreams. “What’s the split on the LEGO sets?” “Sixty-forty. I get the Frozen castle.” “Deal.”
created_at
"2026-02-24T22:16:50.608Z"
2/24/2026, 10:16:50 PM
model
"aion-labs/aion-2.0-20260223"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
2384
moderation_latency
(null)
generation_time
32544
tokens_prompt
77
tokens_completion
1482
native_tokens_prompt
170
native_tokens_completion
1276
native_tokens_completion_images
(null)
native_tokens_reasoning
561
native_tokens_cached
0
num_media_prompt
(null)
num_input_audio_prompt
(null)
num_media_completion
0
num_search_results
(null)
origin
"https://ncbench.com/"
is_byok
false
finish_reason
"stop"
native_finish_reason
"stop"
usage
0.0021776
router
(null)
provider_responses
0
endpoint_id
"1e93b0bb-eb0f-49bc-8fda-947105413b86"
id
"3986054d-db12-4eb6-b010-0508e079b6d8"
is_byok
false
latency
196
model_permaslug
"aion-labs/aion-2.0-20260223"
provider_name
"AionLabs"
status
200
api_type
"completions"
id
"gen-1771971410-6sLGk0gObljjGDbHimFd"
upstream_id
"3986054d-db12-4eb6-b010-0508e079b6d8"
total_cost
0.0021776
cache_discount
(null)
upstream_inference_cost
0
provider_name
"AionLabs"
Evaluation details
Result
Evaluator
Details
Meta Data
0.03%
Matches word count
n/a
neededClean
false
words
530
0.00%
Dialogue to Total Word Ratio
Ratio: 68.72%, Deviation: 18.72%
neededClean
false
wordsTotal
537
wordsDialogue
369
0.0152%