NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-23-01M37PZE0S1P5P1JERZ0HQY2K6
Run Details
Model:
Aion 3.5 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
19.6s
Cost:
$0.005107
Provider:
AionLabs
Evaluation summary
Score
Evaluator
Details Link
51.93%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
25.9628%
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 crouched behind the wheel of a rented van, binoculars pressed to her face, watching the suburban backyard like it was the Louvre. "Okay, I've cased the joint," she said. "One bouncy castle, two adults, approximately fourteen children, and a piñata shaped like a dinosaur." Bill lowered his sunglasses and peered over the dashboard. "That's a lot of kids." "That's a lot of witnesses, Bill. Focus." "I'm focused. I'm just saying, kids are unpredictable. One of them starts crying, they all start crying. It's like dominoes, but loud." "Which is why we go in as the entertainment. I'm the balloon artist. You're my assistant, Sparkles." Bill winced. "Do I have to be Sparkles?" "The costume has a hood. The hood has a bow. You're wearing the bow, Bill." He sighed the sigh of a man who had worn worse bows for smaller scores. "Fine. But last time I was a magician's assistant and a toddler bit me." "That toddler had excellent instincts. Now, the presents are stacked on a table under the maple tree. That's our target. We're talking twelve, maybe fifteen gifts. One of them is shaped like a bike." "A bike? We can't fence a bike." "We're not fencing the bike, Bill. The bike is for my nephew. It's his birthday next week and I refuse to pay retail." "See, that's the kind of planning I respect. What's the exit strategy?" Molly unrolled a hand-drawn map on the dashboard, complete with arrows and a stick figure being launched into the air. "We grab the table, we carry it to the fence line, and we go over the fence into the Hendersons' yard. Their dog is fourteen and blind. We walk through their yard at a leisurely pace, load the van, and we're gone before anyone notices the table is missing." "Kids will notice the table is missing. Kids notice everything. One time a kid noticed I had a tattoo of a shark on my ankle and asked me about it for forty-five minutes." "Then we leave a decoy table. I brought one. It's card table. It's from a garage sale. It cost me two dollars." Bill nodded slowly. "That's good tradecraft." "I know it's good tradecraft. Now, the piñata is the wildcard. If they crack it before we move, we have a candy stampede. Chaos. Which is either our best friend or our worst enemy." "What if we steal the piñata too?" Molly paused, binoculars halfway to her eyes. "Bill. That's the smartest thing you've ever said." "I've been saving up." She started the van and pulled her balloon apron off the seat. "Alright, Sparkles. Hat on. Bow up. Remember, we're professionals." "We're stealing from children, Molly." "We're stealing from children who are going to get more presents at Christmas anyway. Now smile like you love fun." Bill smiled like a man who feared fun, and they stepped out into the sunlight.
created_at
"2026-09-23T18:03:20.493Z"
9/23/2026, 6:03:20 PM
model
"aion-labs/aion-3.5-20260923"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
957
moderation_latency
(null)
generation_time
19555
tokens_prompt
75
tokens_completion
898
native_tokens_prompt
65
native_tokens_completion
843
native_tokens_completion_images
(null)
native_tokens_reasoning
194
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
0.00510675
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.20.0; linux; x64))"
http_referer
(null)
request_id
"req-1790186600-Gw2nvAgaWsdvN091m3ZR"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790186600-m9ybd0GtARpKOEsxYW6o"
upstream_id
"b688319f73cc4a7ead7d4cf293eafc37"
provider_responses
0
endpoint_id
"09b1fc56-3888-42bd-b7a0-926851c7f9d6"
id
"b688319f73cc4a7ead7d4cf293eafc37"
is_byok
false
latency
957
model_permaslug
"aion-labs/aion-3.5-20260923"
provider_name
"AionLabs"
status
200
total_cost
0.00510675
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
51.93%
Matches word count
n/a
neededClean
false
words
484
0.00%
Dialogue to Total Word Ratio
Ratio: 77.57%, Deviation: 27.57%
neededClean
false
wordsTotal
486
wordsDialogue
377
25.9628%