NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-25-01KYCWJ8CJ6NC2Q8FT153SXVKH
Run Details
Model:
Thinking Machines Inkling (Reasoning)
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
4m 56s
Cost:
$0.009343
Provider:
DeepInfra
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 70% 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
Bill tapped the blueprint with a grease-stained finger. "So we're clear? The Harrison kid turns six tomorrow. Thirty guests. Thirty presents minimum." Molly sipped her coffee, black as her heart. "Minimum? Bill, you saw the invite. 'Gifts welcome, but presence is present enough.' That means parents brought double. We're talking a mountain of Lego and dolls that haven't been traumatized yet." "Traumatized?" "By children. There's a window." Bill leaned back in his chair. "Window is guarded by a clown named Bubbles." "Not a real clown. A hired one. With a rainbow wig and questionable juggling ethics." "Exactly. Unpredictable." Molly rolled her eyes. "I'll distract Bubbles. You handle the bouncy castle perimeter." "The castle is deflated by eight." "Then we go at seven fifty-five. The parents will be drunk on prosecco and Pinterest guilt." Bill scratched his chin. "What about the dog? Sprinkles. Pomeranian. Bites ankles with enthusiasm." "I brought bacon-flavored dog treats. Bribe." "You're bribing a dog with bacon treats named bacon?" "It's thematic. Now, entry point: back gate. Code is the birthday boy's birth year, apparently shared on Instagram with a cheerful geotag." "Parents these days." "Exactly. Criminal negligence." Bill nodded slowly. "So we slip in, grab the present table, bag it, vanish before cake." "Not just grab. There's a system. I want the ones with the biggest bows. Big bow equals expensive parental guilt." "You're profiling gifts by ribbon?" "Ribbons don't lie, Bill." He pulled out a notepad. "Disguises. I was thinking caterers." "Caterers wear black. We wear black. We'll blend into the shadows like sophisticated shadows." "Shadow caterers with duffel bags?" Molly grinned. "No duffel. Santa sacks. Festive. If caught, we say we're the entertainment." "You're going to tell parents we're elves?" "Elves steal? No. We'll say we're collecting for charity. The charity of us." Bill wrote that down. "This is either genius or the worst plan since I tried to rob a bank with a water pistol." "That was a squirt gun heist, Bill, and it failed because you laughed." "I was committed to the bit." "Commit to this. The gifts are under the piñata tree near the slide." "Slide is active." "Slide is for children under forty pounds. You're fifty pounds over and I'm... agile." "So you slide?" "I slide with dignity. You take the ladder." Bill looked around the empty room. "There's no ladder." "Then you climb the decorative fence." "I'm allergic to decorative ironwork." "Then stay home and guard the getaway bike." "I'm the driver." "You're the driver who gets lost in cul-de-sacs." "That was one time." "You ended up in a drive-thru." Bill sighed. "Fine. I drive. You slide. We both avoid Bubbles, bribe Sprinkles, sack the bows, and leave before the candles blow out." Molly raised her cup. "To birthday grief and ribbon intelligence." "To not getting grounded by a six-year-old." "Six-year-olds don't have legal authority." "They have parents with phones. One viral video and we're memes." "So be charming on camera. Smile, Bill. You're stealing childhood joy; at least look like you enjoy it." He practiced a grin. "How's this?" "Like a man who just ate a lemon wrapped in glitter." "Perfect. Unsettling but festive." Molly stood, pulling on gloves. "Let's go over the timeline one more time. Seven fifty: gate opens with birthday code. Seven fifty-two: Sprinkles gets bacon. Seven fifty-five: Bubbles gets distracted by my juggling lesson—" "You can't juggle." "I can juggle anxiety and three rubber balls." "That doesn't—" "Seven fifty-eight: present table cleared. Eight: we're on the bike with thirty unwrapped futures." Bill grabbed his coat. "What if there's a security camera?" "I checked. The dad installed a Ring doorbell but aimed it at the driveway, not the yard. Classic mistake. He wants to see deliveries, not prevent theft. We are the delivery now, Bill. The delivery of crime." Bill opened the door. "Molly, this is wrong." "Of course it is. That's why it pays. See you at the gate. Don't wear white. The cake is chocolate." "I always wear black." "Then don't wear cake." They stepped into the rain, two shadows with a sack and a plan ridiculous enough to work.
created_at
"2026-07-25T14:58:43.224Z"
7/25/2026, 2:58:43 PM
model
"thinkingmachines/inkling-20260715"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
996
moderation_latency
(null)
generation_time
296102
tokens_prompt
75
tokens_completion
2592
native_tokens_prompt
75
native_tokens_completion
2295
native_tokens_completion_images
(null)
native_tokens_reasoning
1489
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
0.00934319
router
(null)
provider_responses
0
endpoint_id
"d6b9d68e-115a-49d7-81cd-1ad1c6942b0e"
id
"chatcmpl-RpGVLZIN7rgqOa6rjGO7hAgl"
is_byok
false
latency
72
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-1784991523-N8KnQZazoIFpBzENB2aJ"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1784991523-wKrvniWLeVnHQnN4ARqF"
upstream_id
"chatcmpl-RpGVLZIN7rgqOa6rjGO7hAgl"
total_cost
0.00934319
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
686
0.00%
Dialogue to Total Word Ratio
Ratio: 86.98%, Deviation: 16.98%
neededClean
false
wordsTotal
699
wordsDialogue
608
0.0000%