NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-20-01KHY9CXTN27YYHFWFWG5Q7ZZH
Run Details
Model:
Gemma 3 4B
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
7.6s
Cost:
$0.000058
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 kids’ party. Saturday. Little Timmy Henderson’s fifth birthday. Theme’s ‘Pirates and Rainbows,’ which, frankly, is a logistical nightmare.” “Logistical nightmare is my middle name, Molly. Well, not really. It’s Bartholomew. But you get the point.” “Bartholomew. Seriously? Okay. So, the parents. Heavily distracted, I’m guessing?” “Predictably. Like a flock of pigeons after a dropped crumb. The moms are all vying for the ‘best parent’ award, and the dads… well, the dads are mostly staring at the bouncy castle.” “Excellent. Bouncy castle distraction. Perfect. Now, the presents. We’re talking an estimated…?” “Conservative estimate? Thirty-seven. Optimistic? Sixty-three. I’ve been observing the invitations. They're sent via ‘Insta-Party,’ so everyone's posting pictures of their gift lists. I’ve cross-referenced. It's a treasure trove.” “Thirty-seven to sixty-three. We’ll aim for sixty-three. We need room for bulk wrapping paper. And glitter. Don’t forget the glitter. That’s crucial to the operation.” “Absolutely crucial. It’s the element of chaos, Bill. It’s what makes it *fun*. What about the security?” “The security guard is Reginald. He’s seventy-two, spends most of his time complaining about the humidity, and his eyesight is… let’s just say he’s convinced the balloons are hostile.” “Reginald. Fantastic. So we’re relying on geriatric apathy? This is brilliant.” “It’s efficient. I’ve rigged a distraction – a very loud, operatic rendition of ‘Happy Birthday’ played through a small speaker. He’ll be too busy flinching.” “Operatic ‘Happy Birthday’? Seriously?” “Don’t underestimate the power of a poorly executed soprano, Molly. It deeply unsettles Reginald.” “Okay, okay. Point taken. And the parents? How do we handle the parents?” “We don't. We blend. I’ve procured a pirate hat and a rainbow-colored inflatable parrot. We’ll be ‘eccentric relatives.’ It’s the key. ’Oh, Timmy is such a darling, isn’t he? Just bursting with joy!’” “’Bursting with joy’ while stuffing a Lego pirate ship into a burlap sack. Sounds about right.” “Precision is paramount. We’re not ripping open presents, Molly. We're… curating them. Selecting the most… *deserving* items.” “Deserving? Based on what criteria? Shiny things?” “Generally. And things that make particularly loud noises. Remember the bouncy castle? A drum kit would be ideal.” “Bill, you’re starting to sound like a five-year-old. A very meticulous five-year-old.” “Just ensuring we’re maximizing the chaos, Molly. Maximizing the profit. And the glitter.” “Right. Glitter. Let’s just... aim for sixty-three. And no operatic ‘Happy Birthday’ without my express permission.”
created_at
"2026-02-20T19:46:01.7Z"
2/20/2026, 7:46:01 PM
model
"google/gemma-3-4b-it"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
157
moderation_latency
(null)
generation_time
7391
tokens_prompt
93
tokens_completion
775
native_tokens_prompt
75
native_tokens_completion
693
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
0.00005844
router
(null)
provider_responses
0
endpoint_id
"d3de3bd4-81bc-48fb-924f-2a87b2a36e75"
id
"chatcmpl-RaMABeDjOfUb5KUYb5e0vMI4"
is_byok
false
latency
62
model_permaslug
"google/gemma-3-4b-it"
provider_name
"DeepInfra"
status
200
api_type
"completions"
id
"gen-1771616761-DLVYjYtbC15sEdS3pIkN"
upstream_id
"chatcmpl-RaMABeDjOfUb5KUYb5e0vMI4"
total_cost
0.00005844
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
402
0.00%
Dialogue to Total Word Ratio
Ratio: 100.00%, Deviation: 30.00%
neededClean
false
wordsTotal
402
wordsDialogue
402
0.0000%