NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYCWGTBE3XKKH20Z648QWHN2
Run Details
Model:
Thinking Machines Inkling (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
52.2s
Cost:
$0.025153
Provider:
BaseTen
Evaluation summary
Score
Evaluator
Details Link
0.00%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
0.0000%
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 slammed the van door hard enough to rattle Bill's poorly hidden energy drink. "Remind me why we're targeting a six-year-old's party?" Bill twisted in the passenger seat, his neon green ski mask—sparkly unicorn horn and all—sliding halfway down his face. "Because, Molly, modern children have better gear than special forces. I'm talking a drone with live video, three virtual reality headsets, and allegedly a robotic companion that learns your secret nickname." "Allegedly?" Molly yanked her black beanie over her curls and glared. "We rob banks, Bill. Art galleries. That one collector with the ancient coin hoard and the angry cat. Not bouncy castles." "It's strategic," Bill insisted, unfolding a map stained with coffee and mysterious ketchup. "The parents are distracted by prosecco and panic. There's free cake. And according to the invite I liberated—" "You stole federal mail." "I borrowed it with a very convincing lost-puppy dialogue." He jabbed a finger at a circled address. "The gifts stay in the garage overnight for 'convenient Sunday opening,' according to Mom's social media with forty-seven heart reacts." Molly rubbed her temples. "You study hashtags now?" "I'm interdisciplinary." Bill grinned, horn wobbling. "Forty-seven kids. Twelve adults. One magician named Giggles. And critically—zero security facing the garden gate." "Giggles sounds like a supervillain." "He does card tricks for toddlers, Molly. His nuclear option is making a four-year-old cry over a hidden ace of spades." She pulled her laptop open, fingers flying. "But you're right about the loot. Stacked in the garage like a retail apocalypse. PlayStation Five, motorized jeep, dollhouse with working lights larger than our apartment." Bill leaned closer, his peppermint gum breath hitting her cheek. "Any guards?" "Buttercup?" "Every party has a golden retriever named Buttercup." "I cross-referenced municipal pet licenses. No Buttercup. Tyler's severely allergic, apparently." She snapped the laptop shut. "Real problem is you parked beneath the driveway security dome." Bill squinted through the windshield. "That black eyeball? That's for the swimming pool." "That eyeball records license plates, genius." Molly tossed black gloves into his lap. "Wear these. And ditch the unicorn mask. You're going in as catering staff." Bill held up a white polo with PARTY PALACE misspelled across the chest. "This smells like barbecue tragedy." "Thrift store miracle. Pockets are deep enough for tools." She handed him a cheap black cap. "Pull it low. Your jaw is too memorable." "My jaw gets compliments." "Exactly. You look like a cartoon prince designed by committee." She pushed the van door open. "Here's the clock: Phase one, I plant the signal jammer near their Wi-Fi router. Phase two, at three-fifteen the birthday cake arrives and every child screams like they're being taxed. That's our window. Phase three, side gate, garage, bags, gone." Bill zipped his lips with an imaginary zipper, then instantly broke the seal. "What if Giggles figures us out?" "He'll be too busy pulling plastic flowers from behind ears." Molly hopped onto the pavement, adjusting her backpack. "Ground rules: do not eat the cake, do not liberate the pony—" "There is a pony?" "Read the invite, Bill! Yes, in the yard for exactly thirty minutes. Don't pet it aggressively. Don't steal it." She pointed toward a row of hedges. "Hide in there. And for criminal integrity, do not twist balloon animals inside. Understand?" Bill pulled a rubber chicken from his coat pocket and saluted with dignity. "Operation Birthday Bandit is officially live. But Molly?" "Yeah?" "If this works, I'm framing my cut in a display case with a plaque." Molly shook her head, hiding a smirk. "You're going to ruin everything." "Only if Giggles recognizes my chiseled jawline." He tugged the cap downward. "Wait—what's my cover name?" "Brad. You carry napkins and avoid eye contact." "Brad sounds trustworthy." "Brad sounds like he belongs in the background of someone else's life." She turned to leave, then paused. "Oh, and Bill? Stay out of the bounce castle." "Bounce house." "Whatever. Those nylon prisons are easier to enter than leave. I mean it." Bill watched her vanish into the twilight, then whispered to the rubber chicken, "She worries too much. I've already practiced my balloon giraffe."
created_at
"2026-07-25T14:57:56.087Z"
7/25/2026, 2:57:56 PM
model
"thinkingmachines/inkling-20260715"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
117
moderation_latency
(null)
generation_time
51837
tokens_prompt
75
tokens_completion
6944
native_tokens_prompt
75
native_tokens_completion
6192
native_tokens_completion_images
(null)
native_tokens_reasoning
5233
native_tokens_cached
0
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
0.0251526
router
(null)
provider_responses
0
endpoint_id
"ba2df08c-c0a0-4e0e-9f31-d79a9c475b70"
id
"chatcmpl-17fa5deca38f48a8af8ecccb8d6abfa5"
is_byok
false
latency
117
model_permaslug
"thinkingmachines/inkling-20260715"
provider_name
"BaseTen"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.18.0; linux; x64))"
http_referer
(null)
request_id
"req-1784991476-FTMRko9LL0Aeuc8aWNJc"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1784991476-OLrjeFsShzTnDpwkouma"
upstream_id
"chatcmpl-17fa5deca38f48a8af8ecccb8d6abfa5"
total_cost
0.0251526
cache_discount
(null)
upstream_inference_cost
0
provider_name
"BaseTen"
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
689
0.00%
Dialogue to Total Word Ratio
Ratio: 70.82%, Deviation: 20.82%
neededClean
false
wordsTotal
699
wordsDialogue
495
0.0000%