NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-20-01KHY9XWFPCNERS5ZP6YBZ3C49
Run Details
Model:
Gemma 3 4B
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
6.3s
Cost:
$0.000054
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
“Right, so, the bouncy castle. We’re going in through the air vents.” “Air vents? Seriously, Molly? You’re suggesting we scale a building’s ventilation system to liberate a mountain of plastic toys?” “Don’t be dramatic, Bill. It’s strategically placed. Minimal security. Mostly pigeons. Plus, think of the children!” “I’m thinking of the paperwork, Molly. ‘Attempted infiltration of private residence posing as a child’s birthday party – potential criminal charges’ isn’t a particularly fun reading list.” “Relax. The principal, Mrs. Higgins, is obsessed with gluten-free cupcakes. She’ll be too busy worrying about her vegan frosting to notice a couple of acrobatic professionals.” “You’re relying on a woman who’s currently embroiled in a passive-aggressive campaign against the PTA over homemade sugar cookies. That's hardly a solid security plan." “It’s a distraction! A delicious, sugary distraction! Now, the presents. We’re targeting the plush animals, obviously. Paddington’s a must. And that giant unicorn.” “You want a giant unicorn? Molly, we’re stealing presents for five-year-olds. Not staging a unicorn-themed raid.” “Don’t be pedantic. It’s got sentimental value. Think of the Instagram potential.” “Instagram potential doesn't pay the bail money. Look, let's just focus on getting in and out. No elaborate schemes, no unicorn obsession. Just… presents.” “Fine, fine. But if we miss out on the unicorn, I’m holding you personally responsible for a severe case of existential disappointment.” “You’re impossible. Okay, look, we go in during the cake cutting, right? Chaos. Excitement. Kids are distracted. We grab what we need, quick and quiet.” “Cake cutting? That’s… shockingly ambitious. You're planning a heist during a sugar-fueled frenzy of five-year-olds?” “It’s the optimal window, Bill! Trust me. We use the balloons as cover. Lots of balloons. Big, distracting balloons.” “Balloons? You want to inflate our getaway with helium balloons?” “It’s a method, Bill! A perfectly reasonable, slightly ridiculous method. What’s your brilliant alternative? Sneaking in through the rose bushes?” “The rose bushes are guarded by a chihuahua named Princess Fluffybutt.” “Princess Fluffybutt is a minor detail! We’ll neutralize her. With glitter.” “Glitter? You’re going to attack a chihuahua with glitter?” “It’s a non-lethal deterrent. Highly effective. Look, just agree. Balloons, cake cutting, Paddington, unicorn. Let's just get this done.” “I still think this is profoundly stupid. But… Paddington does look rather dashing.” “See? I knew you'd come around. Now, let's talk disguises…”
created_at
"2026-02-20T19:55:17.379Z"
2/20/2026, 7:55:17 PM
model
"google/gemma-3-4b-it"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
167
moderation_latency
(null)
generation_time
6052
tokens_prompt
93
tokens_completion
744
native_tokens_prompt
75
native_tokens_completion
632
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
0.00005356
router
(null)
provider_responses
0
endpoint_id
"d3de3bd4-81bc-48fb-924f-2a87b2a36e75"
id
"chatcmpl-Rc8iRcQ3kLT6fxkzlKvHTH3q"
is_byok
false
latency
59
model_permaslug
"google/gemma-3-4b-it"
provider_name
"DeepInfra"
status
200
api_type
"completions"
id
"gen-1771617317-dwBFXUTJULNLXkepZupn"
upstream_id
"chatcmpl-Rc8iRcQ3kLT6fxkzlKvHTH3q"
total_cost
0.00005356
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
392
0.00%
Dialogue to Total Word Ratio
Ratio: 100.00%, Deviation: 30.00%
neededClean
false
wordsTotal
392
wordsDialogue
392
0.0000%