NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-20-01KHYA9CVX4ZN3RNBDPF5HXN2Z
Run Details
Model:
Gemma 3 27B
Test:
Dialogue tags
Scenario:
Write 500 words with 50% dialogue
Duration:
14.8s
Cost:
$0.000195
Provider:
Nebius
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 of the diner booth stuck to Molly’s elbows. She traced the floral pattern with a sugar packet, avoiding Bill’s gaze. He was enthusiastically sketching a floor plan on a napkin, using ketchup as ink. “Okay, so the bouncy castle is a definite distraction,” Bill announced, holding up the crimson-stained napkin. “Kids are magnetically drawn to those things. Creates a nice…present-access corridor.” “A corridor? Really, Bill? We’re robbing a six-year-old’s birthday, not infiltrating Fort Knox." “Details, Molly, details. We need a strategy. This isn't just *any* birthday party, this is little Timmy Henderson’s. His parents are loaded. Think solid gold Thomas the Tank Engine, vintage Lego sets…the good stuff.” “You’ve been casing the Henderson’s for a week, haven’t you?” “Research. Thorough research. And I’ve identified a weakness.” “Let me guess. The pinata?” “Almost! The clown.” Bill tapped the napkin with a french fry. “Absolute chaos merchant. He'll be juggling, making balloon animals that vaguely resemble poodles, generally ensuring no one keeps an eye on the loot.” “A clown. Our entire operation hinges on a clown.” Molly sighed, pushing a stray strand of purple hair behind her ear. “This is getting progressively more ridiculous.” “Ridiculously *profitable* is what it is. So, I was thinking, you handle crowd control. Distract the parents with…flattery? Compliment their questionable fashion choices? Pretend you’re a long-lost relative? You’re good at blending.” “Blending? I’m wearing a neon pink jumpsuit, Bill.” “It’s…festive! It’ll fit right in! Think undercover birthday enthusiast. You are a master of improvisation.” “I once convinced a security guard I was a visiting dignitary from Liechtenstein. By wearing a feather boa and speaking with a vaguely European accent.” “Perfect! See? Born for this. I’ll be over by the gift table. Quick in, quick out. Santa sack employed, naturally.” “Santa sack?” Molly raised an eyebrow. “It’s thematic! And large. Very large. Holds approximately forty-seven plush unicorns, I estimate.” “You’ve estimated the capacity of a Santa sack? Bill, I swear…” “Look, preparation is key! What if Timmy gets a disproportionate amount of plushies? We need to be prepared for all eventualities. So, blending, flattery, and a well-timed distraction from the clown. Are we clear?” “Crystal. Just promise me no face paint.” “No promises. It really completes the look.” Bill winked, then went back to his ketchup-stained floorplan, humming a jaunty tune. "Operation: Birthday Bonanza is a go!"
created_at
"2026-02-20T20:01:34.602Z"
2/20/2026, 8:01:34 PM
model
"google/gemma-3-27b-it"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
309
moderation_latency
(null)
generation_time
14725
tokens_prompt
93
tokens_completion
727
native_tokens_prompt
75
native_tokens_completion
625
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
0.000195
router
(null)
provider_responses
0
endpoint_id
"db3af039-8c6e-4b32-bbf5-c858d644d1b5"
id
"chatcmpl-f7bae4c570b14884af2111063ec9fcdb"
is_byok
false
latency
309
model_permaslug
"google/gemma-3-27b-it"
provider_name
"Nebius"
status
200
api_type
"completions"
id
"gen-1771617694-n20SFpWw11lwt8GSVOe7"
upstream_id
"chatcmpl-f7bae4c570b14884af2111063ec9fcdb"
total_cost
0.000195
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Nebius"
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Matches word count
n/a
neededClean
false
words
399
0.00%
Dialogue to Total Word Ratio
Ratio: 79.20%, Deviation: 29.20%
neededClean
false
wordsTotal
399
wordsDialogue
316
0.0000%