NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-20-01KHYB1SRDRS6T7KDT89171G1B
Run Details
Model:
Gemma 3 27B
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
26.4s
Cost:
$0.000114
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 chipped Formica table wobbled under the weight of Bill’s blueprint – a crayon drawing, mostly. It depicted a bouncy castle, several stick figures labelled ‘KIDS’ and, circled repeatedly in red, the word ‘PRESENTS’. “Right,” Molly said, tracing a line with a chipped nail. “So, Operation Stocking Fillers relies heavily on…strategic distraction?” Bill adjusted his glasses, which were perpetually sliding down his nose. “Strategic is a strong word. More like…controlled chaos. Little Timmy’s having a fifth birthday. Five-year-olds are chaos incarnate. We’re just…augmenting it.” “Augmenting. I like that. Sounds less…sinister. So, you’re sticking with the clown disguise?” “Absolutely. It’s foolproof. People expect a clown to be annoying, not a criminal mastermind.” Molly snorted. “You calling yourself a mastermind? Last time you tried a disguise, you ended up looking like a confused badger.” “That was a *very* specific badger,” Bill protested. “And it was effective! No one suspected a thing. Until I tripped over the garden gnome.” “And alerted the entire neighbourhood. Right.” Molly drummed her fingers on the table. “Okay, clown you. I’m going with ‘enthusiastic party parent’. Lots of over-the-top compliments, questions about the cake, the works. Blend in, observe, identify the present stash." “The stash is indicated on the blueprint here,” Bill said, tapping the crayon circle. "Under the bouncy castle. Apparently, Brenda – little Timmy's mother – believes it's a ‘secure location’. Brenda clearly underestimates the allure of inflated plastic.” “Brenda is going to be very disappointed,” Molly agreed cheerfully. “So, timing. The magician starts at two-thirty. That’s our window. The kids will be mesmerized, Brenda will be busy making sure Timmy doesn't eat a rabbit out of a hat, and we swoop.” “Swoop sounds…undignified for a clown.” “Bill, you're wearing oversized shoes and a rainbow wig. Dignity went out the window the moment you bought the honking nose.” “It's a *professional* honking nose!” “Whatever you say. Okay, bags? We're going to need large, inconspicuous bags. I’m thinking reusable shopping totes. Eco-friendly crime.” “Excellent thinking. I’ve already sourced some. They have pictures of kittens on them. Nobody suspects a kitten-tote carrying thief.” “You’re committed to this bit, aren't you?” “Committed is my middle name. Well, actually it's Harold. But I feel committed.” Molly grinned. “Right then, Harold. Let’s go steal some birthday presents.” “Just one quick question,” Bill said, fiddling with his glasses. “What are we going to *do* with all the presents?” “Sell them, obviously. We’re professionals.” “Right. Of course. Professionals. I just…wonder if little Timmy will be upset.”
created_at
"2026-02-20T20:14:54.234Z"
2/20/2026, 8:14:54 PM
model
"google/gemma-3-27b-it"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
156
moderation_latency
(null)
generation_time
26286
tokens_prompt
93
tokens_completion
783
native_tokens_prompt
75
native_tokens_completion
676
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
0.00011416
router
(null)
provider_responses
0
endpoint_id
"8f22002c-c045-446f-a1b9-9896133536b8"
id
"chatcmpl-Rws47IeNgS1Mdf2O3jiXZWnh"
is_byok
false
latency
55
model_permaslug
"google/gemma-3-27b-it"
provider_name
"DeepInfra"
status
200
api_type
"completions"
id
"gen-1771618494-20TNFgsoCIaz5lZCiHB3"
upstream_id
"chatcmpl-Rws47IeNgS1Mdf2O3jiXZWnh"
total_cost
0.00011416
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
417
0.00%
Dialogue to Total Word Ratio
Ratio: 82.73%, Deviation: 32.73%
neededClean
false
wordsTotal
417
wordsDialogue
345
0.0000%