NC Bench
Overview
About
Tests
Models
Vendors
Run: 2026-02-20-01KHYAAV4G0P6MCCRRXNNAQ7T6
Run Details
Model:
Gemma 3 27B
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
21.0s
Cost:
$0.000112
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 lemme get this straight,” Bill said, meticulously polishing his glasses with a slightly grubby handkerchief. “We’re breaking into a *birthday party*?” Molly rolled her eyes, already fiddling with a ridiculous amount of brightly coloured hairpins. “It’s not ‘breaking in’ Bill, it’s ‘strategic present acquisition’. And yes. Seven-year-old Timmy’s going to have a very minimalist birthday.” “Strategic present acquisition? You’re giving this a whole new level of… pretension, even for you, Molly.” “Look, the intel is solid. Timmy’s parents are loaded, they're throwing a party at that bouncy castle place, and apparently, the gift haul is legendary. We're talking a miniature pony's worth of Lego.” Bill blinked. “A miniature pony’s worth? That’s…specific.” “It means a lot, Bill. A *lot*. Enough to finally get that speedboat we talked about.” “The one with the ejector seat?” “The very one! Now, focus. The plan is simple. You create a diversion. Something… distracting.” “Distracting like… juggling chainsaws?” Molly gave him a withering look. “No. Distracting like pretending to be a particularly enthusiastic magician’s assistant. They hired ‘Amazing Al’ which, let's be honest, is setting the bar pretty low. You just need to look helpful and slightly clumsy.” “Helpful and clumsy. That’s basically my default setting.” He sighed. “And what are *you* doing while I’m dodging poorly aimed rabbits?” “I'm going in. The presents are all going to be in the back room, under the guise of ‘safe keeping’. I’ll use the old lock-picking trick, grab the goods, and we’re outta there.” “The ‘old lock-picking trick’ which, last time I checked, involved a bobby pin, a paperclip and a lot of muttered swearing?” “Details, details. It works! Look, timing is everything. Amazing Al does his grand finale - the disappearing hamster - and that’s our cue. Maximum chaos, minimal attention.” “Disappearing hamster… oh, this is getting better and better.” Bill chuckled. "What if the hamster doesn't disappear?" “Then you create a bigger distraction. Pretend to faint. I don’t care. Just buy me time.” “Faint? I haven't fainted since I was twelve and Mrs. Periwinkle made liver and onions." "Bill. Focus. This is for the speedboat.” “Right. The speedboat. But what if Timmy's parents suspect something?" “They won’t. They’ll be too busy wondering how Amazing Al made a hamster vanish into thin air. Besides, who suspects a magician’s assistant and a slightly bewildered-looking gentleman?” “We're going to look ridiculous.” Molly grinned. “Exactly. Totally inconspicuous. It's foolproof, I tell you. Foolproof!”
created_at
"2026-02-20T20:02:21.978Z"
2/20/2026, 8:02:21 PM
model
"google/gemma-3-27b-it"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
440
moderation_latency
(null)
generation_time
20876
tokens_prompt
93
tokens_completion
753
native_tokens_prompt
75
native_tokens_completion
660
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
0.0001116
router
(null)
provider_responses
0
endpoint_id
"8f22002c-c045-446f-a1b9-9896133536b8"
id
"chatcmpl-RJ1aDHrYyxH5UuoHVsJNCTNs"
is_byok
false
latency
44
model_permaslug
"google/gemma-3-27b-it"
provider_name
"DeepInfra"
status
200
api_type
"completions"
id
"gen-1771617741-ve1ZY54OrIwqE0bBmBHJ"
upstream_id
"chatcmpl-RJ1aDHrYyxH5UuoHVsJNCTNs"
total_cost
0.0001116
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
408
0.00%
Dialogue to Total Word Ratio
Ratio: 90.44%, Deviation: 20.44%
neededClean
false
wordsTotal
408
wordsDialogue
369
0.0000%