NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-05-10-01KR8Y19CA33XCTJTEDK1ZT388
Run Details
Model:
Gemini 3.1 Flash Lite
Test:
Dialogue tags
Scenario:
Write 500 words with 70% dialogue
Duration:
3.5s
Cost:
$0.001038
Provider:
Google AI Studio
Evaluation summary
Score
Evaluator
Details Link
95.99%
Matches word count
(details)
82.90%
Dialogue to Total Word Ratio
(details)
89.4427%
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
"Are you entirely certain that they are just toys in those boxes, Bill?" Molly whispered, adjusting her black tactical turtleneck. The fabric itched against her throat like a swarm of angry wool ants. "Positive," Bill replied, fiddling with a pair of oversized clown shoes he insisted would provide traction on the polished gymnasium floor. "High-end electronic kits, remote-controlled helicopters, and at least one limited-edition golden action figure that retails for more than our last two jobs combined." Molly squinted through the sliver of the doorway. Twenty toddlers were currently engaged in a high-stakes scavenger hunt involving plastic eggs. "It is a birthday party, Bill. Why are they guarded by a bouncy castle shaped like a fire-breathing dragon?" "It is a psychological deterrent," Bill said, adjusting his fake red nose. "Parents think it’s festive. I think it’s a tactical obstacle course designed to break our ankles." "Right. And the clown?" "We ignore the clown. If the clown notices us, we tell him we are the backup entertainment. I have a balloon-animal kit in my fanny pack." Molly stared at him. "You practiced making balloon animals for this heist?" "Precision is the hallmark of a professional, Molly. Poodle, sword, maybe a nervous-looking swan if things get hairy." "This is madness. We are casing a room full of juice boxes and sugar-crazed three-year-olds to steal loot that won’t even fit in our getaway van." "It’s not about the size of the loot, it’s about the exclusivity of the haul. That action figure is worth a fortune on the black market for bored collectors." Molly adjusted her goggles. "Fine. How do we bypass the perimeter?" "We wait for the cake-cutting ceremony," Bill whispered, his voice intense. "Chaos threshold peaks when the candles are lit. The parents will be distracted by smartphone cameras and the imminent sugar crash. That is our window." "And the exit strategy?" "We blend in. I’ll make a giraffe, you grab the stack of gifts, and we exit through the side fire door while everyone is singing off-key." "You did not account for the parents," Molly noted, pointing toward a formidable woman in a leopard-print dress checking her watch. "That one has the eyes of a hawk." "She’s just waiting for the goody bags. She won't suspect two adults in clown shoes creeping through the ball pit." Molly sighed and checked her watch. "The cake is coming out. It’s covered in sparklers." "Showtime," Bill said, pulling his rubber nose tight. "Remember, keep your hands off the cupcakes. I don't want to get caught because you have frosting on your tactical gear." "If we end up in jail for a pile of plastic toys, I am documenting exactly how ridiculous this plan was in our memoirs," Molly muttered, stepping out from behind the heavy velvet curtain. "Focus, Molly. The giraffe," Bill hissed. "Stay in character." They crept toward the gift table, the sound of off-key singing rising like a wall of sound around them.
created_at
"2026-05-10T12:33:15.665Z"
5/10/2026, 12:33:15 PM
model
"google/gemini-3.1-flash-lite-20260507"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
594
moderation_latency
(null)
generation_time
3401
tokens_prompt
77
tokens_completion
783
native_tokens_prompt
67
native_tokens_completion
681
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)
num_fetches
(null)
web_search_engine
(null)
origin
"https://ncbench.com/"
is_byok
false
finish_reason
"stop"
native_finish_reason
"STOP"
service_tier
"standard"
usage
0.00103825
router
(null)
provider_responses
0
endpoint_id
"237af107-d2e9-46c7-b146-c74d75980dee"
id
"C3sAasa4M-_Sz7IPzPSkYA"
is_byok
false
latency
594
model_permaslug
"google/gemini-3.1-flash-lite-20260507"
provider_name
"Google AI Studio"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.14.1; linux; x64))"
http_referer
(null)
request_id
"req-1778416395-AZriJqllUgZ2EbZt8Mk3"
session_id
(null)
api_type
"completions"
id
"gen-1778416395-Rq35mODbCHE9ZJH8YKWl"
upstream_id
"C3sAasa4M-_Sz7IPzPSkYA"
total_cost
0.00103825
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Google AI Studio"
response_cache_source_id
(null)
Evaluation details
Result
Evaluator
Details
Meta Data
95.99%
Matches word count
n/a
neededClean
false
words
492
82.90%
Dialogue to Total Word Ratio
Ratio: 71.94%, Deviation: 1.94%
neededClean
false
wordsTotal
506
wordsDialogue
364
89.4427%