NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2025-11-18-01KACP797TWKGQ0FM4X8ZHZX5C
Run Details
Model:
Gemini 2.5 Flash Lite
Test:
Dialogue tags
Scenario:
Write 200 words with 90% dialogue
Duration:
1.5s
Cost:
$0.000123
Provider:
Google AI Studio
Evaluation summary
Score
Evaluator
Details Link
68.10%
Matches word count
(details)
100.00%
Dialogue to Total Word Ratio
(details)
84.0500%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 200 word scene that has 90% 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, Bill, you’re sure about this?” Molly whispered, adjusting the fake unicorn horn perched precariously on her head. “Positive. The intel is solid. Free balloons, sugar rushes, and a veritable smorgasbord of… loot,” Bill replied, fumbling with a bright pink noisemaker. “Loot? It’s just toys, Bill. And a lot of sticky little fingers.” “Precisely! The perfect cover. Who’s going to suspect the guy with the suspiciously well-stuffed teddy bear costume of anything untoward?” “You’re wearing a teddy bear costume? I thought we agreed on the clown thing.” “Clowns are too intimidating for kids, Molly. Bears are cuddly. Plus, it helps conceal the crowbar.” “Bill, I can’t believe we’re doing this. Stealing presents from toddlers.” “Think of it as an… early holiday redistribution. Besides, their parents probably bought them too much anyway. We’re providing a valuable service. Now, you distract the bouncy castle. I’ll handle the gift table. Remember, stealth is key. And try not to inhale glitter this time.” “No promises. Just make sure you don’t accidentally hug a child when you go for the big G.I. Joe.” “Duly noted. Operation Joyful pilferage is a go.”
created_at
"2025-11-18T23:53:57.38905+00:00"
11/18/2025, 11:53:57 PM
model
"google/gemini-2.5-flash-lite"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
288
moderation_latency
(null)
generation_time
1413
tokens_prompt
68
tokens_completion
274
native_tokens_prompt
67
native_tokens_completion
290
native_tokens_completion_images
0
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
0.0001227
api_type
"completions"
id
"gen-1763510035-EEX7jtIs9Ak5ZOlhleoD"
upstream_id
(null)
total_cost
0.0001227
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Google AI Studio"
Evaluation details
Result
Evaluator
Details
Meta Data
68.10%
Matches word count
n/a
neededClean
false
words
186
100.00%
Dialogue to Total Word Ratio
Ratio: 89.30%, Deviation: 0.70%
neededClean
false
wordsTotal
187
wordsDialogue
167
84.0500%