NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-09-23-01M37NXWSPQG7ZNYC67E907ED6
Run Details
Model:
Aion 3.5 (Reasoning, High)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
18.4s
Cost:
$0.004747
Provider:
AionLabs
Evaluation summary
Score
Evaluator
Details Link
75.16%
Matches word count
(details)
94.46%
Dialogue to Total Word Ratio
(details)
84.8062%
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 a hydrangea bush and studied the suburban backyard like it was the vault at Fort Knox. Through the fence slats, she could see the target: a folding table groaning under the weight of presents, wrapped in paper emblazoned with dinosaurs wearing party hats. "There it is," she whispered. "The payload." Bill lowered his binoculars, which was unnecessary, since they were eleven feet away. "I count fourteen packages. Maybe fifteen. There's something oblong behind the bike." "Focus, Bill." They had been planning this for three days, ever since Molly spotted the party invitation taped to a neighbor's refrigerator during a completely unrelated burglary. The plan was elegant in its simplicity: enter through the side gate, load the presents into a wagon, exit before anyone noticed. Bill had drawn a diagram on a napkin. It was mostly a square with an arrow. "Here's the thing," Bill said, chewing thoughtfully on a granola bar he'd brought for morale. "I've been doing recon on the perimeter, and I've identified the primary threat." "The clown?" "The clown is a secondary threat. The primary threat is a child named Tyler. He's the birthday boy, he's turning seven, and according to intel from the bounce house, he bites." "Intel from the bounce house?" "I stood near the bounce house and listened, Molly. That's how intelligence gathering works." Molly rubbed her temples. Somewhere beyond the fence, a shriek rang out, followed by the unmistakable sound of a piƱata being beaten to death. "Change of plan," she said. "We go in during the cake. When they sing, we move. Nobody watches the presents during 'Happy Birthday.' It's a known blind spot in party security." "That's actually genius," Bill admitted. "I read about it in a magazine." "You didn't read that anywhere." The singing began at 3:14 p.m. Bill threw open the gate with all the subtlety of a fire alarm, and they sprinted for the table, Molly scooping presents into a pillowcase while Bill pulled the wagon. It was going perfectly. It was going magnificently. Then Tyler appeared directly in front of them, holding a juice box, his eyes narrowing with the suspicion of a seasoned customs officer. "You're stealing the presents," he said flatly. "No," said Bill, sweating through his disguise, which was a fake mustache he'd forgotten to put on. "We're relocating them. For safety." Tyler bit Bill on the ankle. By the time they cleared the fence, the pillowcase held two presents and one partially eaten juice box, and Bill was limping and bleeding and oddly proud. "Two out of fifteen," Molly panted, as sirens that were definitely not for them but felt symbolic echoed somewhere. "Terrible yield. Abysmal." Bill held up the larger box and shook it near his ear. "It's a robot. Molly. It's a robot kit. This is a win." It was not a robot kit. It was a candle-making set, which they would later agree was the universe's way of mocking them, as they sat in their van, sulking, making a single lavender-scented candle together by the light of a dashboard bulb.
created_at
"2026-09-23T17:45:01.512Z"
9/23/2026, 5:45:01 PM
model
"aion-labs/aion-3.5-20260923"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
424
moderation_latency
(null)
generation_time
18367
tokens_prompt
75
tokens_completion
894
native_tokens_prompt
65
native_tokens_completion
783
native_tokens_completion_images
(null)
native_tokens_reasoning
115
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
0.00474675
router
(null)
user_agent
"langchainjs-openai/1.0.0 ((node/v24.20.0; linux; x64))"
http_referer
(null)
request_id
"req-1790185501-DRYSRdjVKnBzXOkMS2GY"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1790185501-UNJ9LktyPbNO3EMAwGe0"
upstream_id
"e06d5a0872b748ab95754d2d0f024d96"
provider_responses
0
endpoint_id
"09b1fc56-3888-42bd-b7a0-926851c7f9d6"
id
"e06d5a0872b748ab95754d2d0f024d96"
is_byok
false
latency
424
model_permaslug
"aion-labs/aion-3.5-20260923"
provider_name
"AionLabs"
status
200
total_cost
0.00474675
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
75.16%
Matches word count
n/a
neededClean
false
words
513
94.46%
Dialogue to Total Word Ratio
Ratio: 31.07%, Deviation: 1.07%
neededClean
false
wordsTotal
515
wordsDialogue
160
84.8062%