NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-03-03-01KJTYPQ4Z4V1ARHGC1FJ0GDDC
Run Details
Model:
Gemini 3.1 Flash Lite (Preview)
Test:
Dialogue tags
Scenario:
Write 500 words with 30% dialogue
Duration:
3.3s
Cost:
$0.000912
Provider:
Google AI Studio
Evaluation summary
Score
Evaluator
Details Link
2.01%
Matches word count
(details)
0.00%
Dialogue to Total Word Ratio
(details)
1.0058%
Overall
Model input
User/Human
You are a author of comedic heist books. Please write a 500 word scene that has 30% 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
Molly adjusted her neon-pink clown nose, the foam pinching her nostrils until she could barely inhale. Beside her, Bill was wrestling with an oversized balloon animal poodle that kept popping him lightly in the ear. They crouched behind a manicured hedge in the backyard of the Miller estate, their dark tactical gear clashing violently with the festive streamers fluttering in the suburban breeze. The backyard was a minefield of inflatable castles and screaming toddlers high on raw sugar. Are those laser sensors or just fairy lights draped over the fence, Molly whispered, squinting at a rhythmic blink of red LEDs. Bill leaned in, his heavy tactical vest groaning under the weight of his lock-picking kit. Those are definitely fairy lights, but the giant glittery gate is guarded by a golden retriever with the eyes of a professional assassin. We don't need to fight the dog, Molly hissed, shifting her weight as a rogue beach ball collided with her knee. We just need the loot. Did you secure the blueprints for the gift table? I have them right here, Bill said, pulling out a crumpled napkin stained with frosting. The target is located under the primary gazebo. It is guarded by a woman in a tutu who seems to be judging everyone’s parenting style. That is the mother, Molly noted, narrowing her eyes. Is she armed? She is wielding a professional-grade camera and potentially a very sharp cake knife, Bill replied. Perfect, Molly grinned, adjusting her oversized polka-dot gloves. When I deploy the distraction, you grab the haul. I want the pile in the corner that looks like it contains the high-end gaming consoles. How are you planning to handle a distraction, Bill asked, looking skeptically at the rubber chicken hanging from her utility belt. Don't tell me you are going to use the avian poultry decoy again. It failed spectacularly at the museum job. This is different, Molly insisted. That was a serious gala. These are children. They have the attention span of a goldfish. I am going to sprint through the center of the yard screaming that the ice cream truck is being held for ransom. That is the stupidest plan I have ever heard, Bill sighed, though he began checking his bag for extra garbage sacks. They are going to chase us until we hit the state line. Exactly, Molly said, checking her watch. We need them agitated. Agitation leads to chaos, and chaos is our best friend. Move on my mark. Three, two, one, go. Bill watched as Molly vaulted over the hedge, landing in a graceful roll before popping up to perform an unhinged interpretive dance. She let out a piercing shriek about melting popsicles, and the back lawn erupted into absolute pandemonium. Bill lunged toward the gazebo, blending into the confusion, a silent shadow among a whirlwind of sticky-fingered chaos.
created_at
"2026-03-03T22:57:06.736Z"
3/3/2026, 10:57:06 PM
model
"google/gemini-3.1-flash-lite-preview-20260303"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
597
moderation_latency
(null)
generation_time
3179
tokens_prompt
77
tokens_completion
722
native_tokens_prompt
67
native_tokens_completion
597
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
0.00091225
router
(null)
provider_responses
0
endpoint_id
"e004ce39-eb0c-4474-93d5-55072e725970"
is_byok
false
latency
597
model_permaslug
"google/gemini-3.1-flash-lite-preview-20260303"
provider_name
"Google AI Studio"
status
200
api_type
"completions"
id
"gen-1772578626-4pH6z5zkWqpHEt0e7kuG"
upstream_id
(null)
total_cost
0.00091225
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Google AI Studio"
Evaluation details
Result
Evaluator
Details
Meta Data
2.01%
Matches word count
n/a
neededClean
false
words
475
0.00%
Dialogue to Total Word Ratio
Ratio: 0.00%, Deviation: 30.00%
neededClean
false
wordsTotal
481
wordsDialogue
0
1.0058%