NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-20-01KHYASPR7NA0NBA25T2YRMGDZ
Run Details
Model:
Gemma 3 12B
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
14.5s
Cost:
$0.000095
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 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
The air in Molly’s garage smelled overwhelmingly of stale donuts and desperation. Bill, a man whose mustache seemed to defy gravity, was pacing a groove in the concrete floor. Molly, meanwhile, was meticulously sorting miniature plastic ponies into color categories. “Ponies, Molly? Really? We're planning a heist, a geriatric-level audacious heist, and you’re alphabetizing equine figurines?” Bill stopped pacing and threw his hands up. “Strategic reconnaissance, Bill,” Molly said without looking up. “Kids love ponies. Knowing their preferences informs our…camouflage strategy.” “Camouflage. Right. We're going to blend in with a horde of sugar-fueled seven-year-olds dressed as pirates and princesses. Brilliant.” Bill’s mustache quivered. “Look, I'm just saying, are we *sure* about this? Stealing presents? It's practically stealing Christmas, only…earlier and with more glitter.” Molly finally looked up, her expression a mixture of exasperation and amusement. "Bill, we’ve discussed this. The Von Hess estate year-round security is tighter than a drum. Their jewelry collection? Fort Knox. We needed something low-risk, high-reward, and frankly, pathetic enough that the authorities wouldn’t devote significant resources. A children’s birthday party. It’s practically begging to be robbed.” “Pathetic?” Bill sputtered. “They’re kids! They'll cry! What about the parents? They'll be furious!” “They’ll be angry, sure,” Molly conceded. “But anger doesn’t warrant an FBI investigation. A few missing Lego sets? ‘Oh, little Tiffany can’t find her My Little Pony castle. How unfortunate.’ Case closed.” “But what's the plan, exactly?” Bill asked, begrudgingly. “Besides you wearing a tiara and pretending to be a fairy godmother?” “Plan is simple. We go in disguised as… party entertainers. Slightly eccentric, definitely harmless. You’ll be ‘Professor Bumble,’ renowned expert in balloon animals.” Bill gaped. "Balloon animals? I can barely blow up a regular balloon!" “Practice,” Molly said, shoving a crumpled instruction manual at him. “There's a particularly impressive poodle tutorial. Think of the chaos, Bill! While you’re distracting them with dachshunds and swans, I’ll be ‘Sparkle the Magician,’ charming everyone else and liberating the loot.” “Liberating? You mean stealing. And what exactly constitutes ‘loot’?” Bill asked, squinting at the poodle diagram. “Mostly high-end toys, action figures, designer crafts. Mrs. Von Hess doesn’t skimp on presents, apparently. We're talking potential for a small fortune in plastic and plush.” Molly’s eyes gleamed. “Plus, think of the viral videos! ‘Eccentric Balloon Artist Foiled in Birthday Party Heist!’” Bill considered this. "Viral videos..." He mumbled, tentatively blowing into a balloon. It deflated with a pathetic wheeze. "Okay, maybe there's something to this. But I’m drawing the line at dressing up as a unicorn.” “Don't push your luck," Molly warned, returning to her ponies. “Now, tell me, Bill, what’s the most humiliating balloon animal you can conjure?”
created_at
"2026-02-20T20:10:29.009Z"
2/20/2026, 8:10:29 PM
model
"google/gemma-3-12b-it"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
431
moderation_latency
(null)
generation_time
14417
tokens_prompt
93
tokens_completion
837
native_tokens_prompt
75
native_tokens_completion
704
native_tokens_completion_images
(null)
native_tokens_reasoning
0
native_tokens_cached
0
num_media_prompt
(null)
num_input_audio_prompt
(null)
num_media_completion
0
num_search_results
(null)
origin
"https://ncbench.com/"
is_byok
false
finish_reason
"stop"
native_finish_reason
"stop"
usage
0.00009452
router
(null)
provider_responses
0
endpoint_id
"eb06dc92-5a16-47ec-a776-6ef956457c47"
id
"chatcmpl-RkoB5K5Hqb8IU28ZrLzVcKQr"
is_byok
false
latency
107
model_permaslug
"google/gemma-3-12b-it"
provider_name
"DeepInfra"
status
200
api_type
"completions"
id
"gen-1771618229-DE7eiZsTrjxuL12avbJV"
upstream_id
"chatcmpl-RkoB5K5Hqb8IU28ZrLzVcKQr"
total_cost
0.00009452
cache_discount
(null)
upstream_inference_cost
0
provider_name
"DeepInfra"
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
448
0.00%
Dialogue to Total Word Ratio
Ratio: 73.66%, Deviation: 23.66%
neededClean
false
wordsTotal
448
wordsDialogue
330
0.0000%