NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYCSEZJAGGBGRGWAY38RFRA8
Run Details
Model:
Thinking Machines Inkling (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
12m 28s
Cost:
$0.044165
Provider:
DeepInfra
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
98.00%
Dialogue to Total Word Ratio
(details)
48.9982%
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
The van smelled of expired pepperoni and decisions that would eventually require expensive therapy. Its paint job, a shade best described as "regret green," blended poorly with the manicured lawns of suburbia. Bill sat shotgun, cradling a duffel bag embroidered with sparkly unicorns—Molly’s contribution to tactical camouflage and her personal vendetta against conventional taste. “We’re really doing this?” Bill asked, adjusting a collar that suddenly felt like a noose. Molly adjusted the rearview mirror without glancing at it. “Professionals, Bill. This is Tuesday with a piñata.” Outside, 42 Maple Drive blazed like a birthday candle left too near the curtains. Streamers snapped in the breeze. A bouncy castle groaned in the yard like an inflatable beast suffering an existential crisis. The moon watched, indifferent and full, like a judge who had already read the verdict. Bill peeled open a granola bar that had expired during the previous presidential administration. “It’s a child’s party. Cake. Balloons. Moral high ground.” Molly turned, eyes sharp behind sunglasses purchased from a gas station at three in the morning. “High ground doesn’t pay rent. My source says the grandparents are airlifting gifts—electronics, bicycles, a drone that costs more than this van and probably requires a license we don’t have.” “You want to replace Santa with grand larceny.” “I want to diversify our portfolio before we retire into a dumpster.” Bill sighed. The granola tasted of regret, oats, and broken promises. “Entry plan? Front door? Through the cake?” “Service entrance,” Molly said, tapping a laminated floor plan with a chipped fingernail. “Kitchen window. Left unlocked because parents never learn. We slide in during the pin-the-tail chaos—everyone blindfolded, dignity already stolen. We grab the present pile, bag it, and vanish before the candles go out. Fifteen minutes, in and out, like surgeons with worse ethics.” Bill studied the unicorn bag glowing softly in the dark. “This is humiliating.” “It’s tactical misdirection. A black duffel screams felony. Sparkles scream invited guest with excellent manners.” Bill zipped his jacket against the chill of his own conscience. His heart hammered like a toddler who had discovered espresso. “If we get caught, this was your idea.” Molly pushed open the sliding door. The night air arrived sweet with frosting and poor judgment. “Ready?” “As I’ll ever be,” Bill said. They crossed the lawn like shadows wearing questionable fashion, passing a tower of wrapped boxes visible through the bay window. Molly picked the lock with a hairpin she’d liberated from a clown earlier—pure professionalism, questionable hygiene. Inside, chaos reigned with glittering authority. Children shrieked in octaves unknown to science. A magician failed to conjure a rabbit and instead produced a parking ticket with impressive sleight of hand. Parents circled the gift table with the intensity of dragons protecting a particularly shiny hoard. Molly nudged Bill’s ribs. “Distraction in three, two—” Bill sneezed with theatrical violence. Every head turned. Molly smiled, sharp and unrepentant. “Not that kind of distraction.” But Bill was already weaving through the crowd, unicorn bag swinging, with the grace of a man who had definitely rehearsed in a mirror. Molly followed, her hands swift, her conscience lighter than the loot she was about to carry. The presents disappeared into the sparkly void one by one—ribboned, undeserved, utterly stolen. Bill whispered, “We’re monsters.” Molly whispered back, “Monsters with excellent time management.” Then the lights flickered. Someone screamed that the cake had achieved sentience and was on fire. Molly grabbed Bill’s collar. “That’s our exit music.” They bolted for the kitchen window, the unicorn bag bouncing like an excited accomplice, hearts racing, morality firmly abandoned with the wrapping paper and whatever dignity they had brought with them.
created_at
"2026-07-25T14:04:30.163Z"
7/25/2026, 2:04:30 PM
model
"thinkingmachines/inkling-20260715"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
343
moderation_latency
(null)
generation_time
747534
tokens_prompt
75
tokens_completion
11176
native_tokens_prompt
75
native_tokens_completion
10893
native_tokens_completion_images
(null)
native_tokens_reasoning
10127
native_tokens_cached
32
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
0.04416509
router
(null)
provider_responses
0
endpoint_id
"d6b9d68e-115a-49d7-81cd-1ad1c6942b0e"
id
"chatcmpl-RjteQiMtVEwbn6kRvgeWKY1B"
is_byok
false
latency
71
model_permaslug
"thinkingmachines/inkling-20260715"
provider_name
"DeepInfra"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.18.0; linux; x64))"
http_referer
(null)
request_id
"req-1784988270-lCLWqUI7KgUUcBbbiXUg"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1784988270-tGIlYyJNBcSdwxntwftQ"
upstream_id
"chatcmpl-RjteQiMtVEwbn6kRvgeWKY1B"
total_cost
0.04416509
cache_discount
0.00002656
upstream_inference_cost
0
provider_name
"DeepInfra"
response_cache_source_id
(null)
data_region
"global"
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
611
98.00%
Dialogue to Total Word Ratio
Ratio: 29.36%, Deviation: 0.64%
neededClean
false
wordsTotal
613
wordsDialogue
180
48.9982%